You are on page 1of 10

The Laryngoscope

© 2019 The American Laryngological,


Rhinological and Otological Society, Inc.

Systematic Review

Factors of Success of Low-Dose Macrolides in Chronic


Sinusitis: Systematic Review and Meta-analysis

Kachorn Seresirikachorn, MD ; Nichana Suwanparin; Chanyanuch Srisunthornphanich;


Wirach Chitsuthipakorn, MD ; Dichapong Kanjanawasee, MD ; Kornkiat Snidvongs, MD, PhD

Objective: To assess the prognostic factors that predict favorable outcomes of low-dose macrolides (LDMs) in treating
chronic rhinosinusitis (CRS).
Methods: Randomized controlled trials studying the effects of LDMs in treating CRS were included. Data were pooled for
meta-analysis. Primary outcome was Sino-Nasal Outcome Test (SNOT). Six prognostic factors: CRS subtypes, serum immuno-
globulin (Ig)E level, membered lactone ring of macrolides, concurrent endoscopic sinus surgery (ESS), and dosage and duration
of the LDMs were assessed by subgroup analyses.
Results: Ten studies (608 patients) met the inclusion criteria. LDMs and placebo were not different in SNOT improvement
(standardized mean difference [SMD] = −0.23, 95% confidence interval [CI]: −0.69 to 0.24). Subgroup analyses showed that the
effects favored LDMs in the patients with CRS without polyps (SMD = −0.64, 95% CI: −1.01 to −0.27) compared to CRS with polyps,
and the patients receiving a half dose (SMD = −0.64, 95% CI: −1.01 to −0.27) compared to a very low dose. There was no difference
in SNOT improvement between LDMs plus standard treatment compared to standard treatment (SMD = −0.52, 95% CI: −1.57 to
0.53). Subgroup analyses showed that the effects favored LDMs in the patients receiving LDMs for a duration of 24 weeks
(SMD = −1.68, 95% CI: −2.40 to −0.95) compared to 8 and 12 weeks. There was no difference between the 14-membered and
15-membered ring LDMs. Assessment of concurrent ESS found mixed results. Serum IgE level could not be assessed.
Conclusions: LDMs provided favorable outcomes in patients with CRS without polyps. A half dose of macrolides should
be given for a duration of 24 weeks.
Key Words: Macrolides, sinusitis, nasal polyps, anti-inflammatory, dose-response relationship, intranasal surgery.
Level of Evidence: 1a
Laryngoscope, 9999:1–10, 2019

INTRODUCTION agents are the medical treatment for the long-term control
Chronic rhinosinusitis (CRS) with nasal polyps (CRSwNP) of a variety of chronic inflammations.
and CRS without nasal polyps (CRSsNP) represent multi- Macrolides are acknowledged anti-inflammatory agents
ple overlapping entities with various inflammatory pat- with immunomodulatory effects.4,5 They modulate neutro-
terns.1 Diversity in the pathogenesis of CRS associates philic action by suppression of lipopolysaccharide-induced
with a wide spectrum of immunologic profiles and expres- neutrophil migration. The production of proinflammatory
sions of various T helper (Th) cell types and simultaneous cytokines, such as interleukin (IL)-8 and tumor necrosis
expression of multiple Th cell types has been shown in factor-alpha (TNF-α) is suppressed.6 In addition, they modu-
some patients.2 Thus, heterogeneity among patients with late the synthesis and secretion of mucus and alter mucus
CRS is so remarkable that patients may be classified into rheological properties, which results in effective clearance.6
10 clusters by characteristic cytokines.3 Anti-inflammatory Low-dose macrolides (LDMs) have been commonly utilized
for treating upper airway diseases after its clinical effective-
From the Department of Otolaryngology (K.S., D.K., K.S.); Faculty of
Medicine (N.S., C.S.), Chulalongkorn University, Bangkok, Thailand;
ness on diffuse panbronchiolitis was revealed.7
Endoscopic Nasal and Sinus Surgery Excellence Center (K.S., D.K., K.S.), Currently, long-term LDMs therapy in the manage-
King Chulalongkorn Memorial Hospital, Bangkok, Thailand; and the ment of CRS is controversial. Although recommended by
Department of Otolaryngology (W.C.), Rajavithi Hospital Hospital, Bangkok,
Thailand. international guidelines, the evidence supporting LDMs
Editor’s Note: This Manuscript was accepted for publication on therapy is mixed. The first randomized controlled trial (RCT)
January 22, 2019. showed clinical improvement,8 but another RCT showed
K.S. received honoraria for speaking at symposia for Merck Sharp
and Dohme, Mylan, and Menarini. no difference.9 In addition, there is no consensus among
The authors have no other funding, financial relationships, or con- international guidelines on patient selection. The European
flicts of interest to disclose.
Send correspondence to Kornkiat Snidvongs, MD, Department of Otolar- Position Paper on Rhinosinusitis and Nasal Polyps 2012
yngology, Faculty of Medicine, Chulalongkorn University, 1873 Rama 4 Road, (EPOS 2012)1 and the International Consensus Statement
Pathumwan, Bangkok 10330, Thailand. E-mail: drkornkiat@yahoo.com
on Allergy and Rhinology10 recommend macrolides for both
DOI: 10.1002/lary.27865 CRSsNP and CRSwNP subtypes. However, a systematic

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


1
review by Rudmik and Soler11 recommended macrolides data collection form. Six prognostic factors were collected including
for patients with CRSsNP but were against macrolides CRS subtype, serum IgE level, surgical status, membered lactone
for patients with CRSwNP. Furthermore, Haxel et al.12 ring of macrolides, dose, and duration of treatment. The outcomes
were collected at the end of the treatment. A change from the base-
reported no differences between the patients with low and
line with a standard deviation was extracted. A final score was
high serum IgE, although the EPOS 2012 suggested that extracted when a change from the baseline was not reported.13
patients with low serum IgE were macrolide responders. Standard error, interquartile range and 95% confidence intervals
RCTs studying clinical effectiveness of LDMs are heteroge- (95% CI) were used when a standard deviation was not reported.13
nous. The participants were different, not only in CRS sub- Durability of outcomes improvement was collected if available.
types (CRSwNP or CRSsNP) but also surgical status
(without or with surgery). The types (14-membered lactone
Risk of Bias in Individual Studies
ring macrolides and 15-membered lactone ring macrolides), The risk of bias of the included studies was assessed according
the dosages (half dose and very low dose), and the duration of to the Cochrane Handbook for Systematic Reviews of Interven-
treatment (less or longer than 12 weeks) of LDMs varied. We tions.13 Five domains were assessed: 1) random sequence genera-
hypothesized that the anti-inflammatory and immunomodu- tion, 2) allocation concealment, 3) blinding of outcome assessment,
latory effects of macrolides at optimal regimens should be 4) incomplete outcome data, and 5) selective reporting.
effective for specific subgroups. This study aimed to assess
the prognostic factors of LDMs therapy that may predict the Data Synthesis and Statistical Analysis
favorable clinical outcomes by performing a meta-analysis Data were pooled for the meta-analysis. The mean difference
and subgroup analyses. (MD), standard mean difference (SMD) and 95% confidence interval
(CI) were used for continuous data. The heterogeneity (I2) was used
to assess the discrepancies in the treatment effects between differ-
MATERIALS AND METHODS ent trials. An I2 of <40%, 40% to 60%, and > 60% represented low,
This systematic review followed the Preferred Reporting moderate, and substantial heterogeneity, respectively. A fixed-
Items for Systematic reviews and Meta-Analyses statement. effects method was used when the statistical heterogeneity was low.
A random-effects method was used when the statistical heterogene-
ity was high, to provide a more conservative estimate of the differ-
Eligibility Criteria
ences. Statistical assessments were performed using Review
RCTs studying the effects of LDMs therapy on patients with
Manager (RevMan) version 5.3 (The Nordic Cochrane Centre, The
CRS were screened. The diagnostic criteria of CRS depended on
Cochrane Collaboration, Copenhagen, Denmark).14
individual studies. The inclusion criteria included the patients diag-
nosed with CRS by the study authors. The patients were 18 years
old or older. Any type, dosage, and the LDMs given either after
endoscopic sinus surgery (ESS) or without ESS were included. Any Subgroup Analysis, Meta-regression, and
cointerventions were allowed if they were given to both arms of the Sensitivity Analysis
study. The comparisons were 1) LDMs versus placebo, 2) LDMs plus Subgroup analyses by six prognostic factors were assessed.
standard treatment versus standard treatment, and 3) LDMs There were two subgroups of the CRS subtype: CRSwNP and
versus standard treatments. The outcomes were the Sino-Nasal CRSsNP. Studies with a mixed population were assigned to either
Outcome Test (SNOT), symptom score, computed tomography CRSsNP or CRSwNP according to the majority of the population.
(CT) score, endoscopy score, and gastrointestinal and cardiac There were two subgroups of the serum IgE level: high serum IgE
adverse effects. Studies were excluded if the LDMs were given for a (>100 IU/mL) and low serum IgE. There were two subgroups of
short-term duration of less than 6 weeks. The published RCTs in a the concurrent ESS: macrolides given with and without ESS.
language other than English were excluded. There were two subgroups of the dosage: half dose and less than
half dose. The half doses of LDMs were defined as roxithromycin
150 mg once daily (OD), clarithromycin 500 mg OD, azithromycin
Information Sources and Search Strategy 250 mg OD, and erythromycin 500 mg OD. There were two sub-
MEDLINE and Embase were searched using the terms: groups of the membered lactone ring: 14-membered lactone ring
“sinusitis OR rhinosinusitis OR nasal polyp OR sinus surgery” and 15-membered lactone ring. There were three subgroups of the
AND “macrolide OR erythromycin OR clarithromycin OR roxithro- duration of treatment: <12 weeks, 12 weeks, and longer than
mycin OR azithromycin.” The last search was performed on March 12 weeks. Sensitivity analysis was performed by excluding studies
17, 2018. References of the included studies and additional with a high risk of bias at more than one domain. Meta-regression
sources were searched to identify any published or unpublished analyses would be performed in case the number of included stud-
trials that were missed. ies was greater >10.

Study Selection Process


The RCTs selection was performed independently by two RESULTS
reviewers (N.S. and C.S.). The reviewers independently screened
the titles and abstracts based on the predetermined eligibility Study Selection
criteria. The full texts of the selected articles were reviewed. Any A total of 3,301 RCTs were screened (3,299 studies from
disagreements were resolved by consulting the corresponding electronic searches and two studies from manual searches).
author (K.SNIDVONGS), if necessary. After screening, 3,269 studies were excluded. Twenty-two
studies were removed after the full-text review. The reasons
Data Extraction for exclusion are listed in Figure 1. Finally, 10 studies were
Two review authors (K.SERESIRIKACHORN, W.C.) independently included for qualitative synthesis8,9,12,15–21 and nine studies
extracted data from the included studies using a predetermined for quantitative synthesis.8,9,12,15,16,18–21 Characteristics of

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


2
the included studies are shown in Table I. A flowchart of the used azithromycin.9,15 Four trials used a half dose,8,15,17,19
study retrieval and selection is presented in Figure 1. and six trials used less than a half dose.9,12,16,18,20,21
Six trials gave LDMs without ESS.8,9,16–18,21 Of the
four trials giving LDMs with concurrent ESS,12,15,19,20
Participants one trial gave preoperative LDMs19 and three trials gave
Ten trials studied 608 participants, 50.3% were male, postoperative ESS.12,15,20
with the mean age of 43.9 years (nine studies9,12,15–21). All Three trials gave LDMs with a duration of 8 weeks,17–19
patients were adults with CRS: CRSsNP (three trials8,17,21), and six trials for 12 weeks.8,9,12,15,16,21 One trial gave LDMs
CRSwNP (three trials18–20), and mixed subtypes of CRS to one arm with a duration of 12 weeks with a 36% dropout
(a major population of CRSwNP [three trials9,12,16] and rate, and another arm for 24 weeks with no dropouts.20 The
CRSsNP [one trial15]). Two trials measured the serum IgE 24-week duration was included for data extraction and
level at enrollment.8,12 Both studies had mixed populations analysis.
of low and high serum IgE. Six trials gave concomitant medication. Of the six trials,
two trials gave concomitant nasal saline irrigation and intra-
nasal steroid spray,12,15 three trials gave intranasal steroid
Intervention spray,16,18,20 and one trial gave nasal saline irrigation.9
Eight trials assessed the effects of 14-membered lactone
ring macrolides.8,12,16–21 Of the eight trials, five trials used Comparisons
clarithromycin,16,18–21 two trials used erythromycin,12,17 and Four trials compared LDMs therapy versus placebo.8,9,12,15
one trial used roxithromycin.8 Two trials assessed the effects Four trials compared LDMs therapy plus standard treat-
of a 15-membered lactone ring macrolide, and both trials ment versus standard treatment.16,18–20 One trial compared

Fig. 1. Flowcart of the study selection.

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


3
TABLE I.
Characteristics of Included Studies.
CRS Concurrent No. of No. of No. of Duration of
First Author Year Subtype ESS Patients Macrolides Control Macrolides Dose (mg/d) Control Treatment (wk)

Wallwork 2006 CRSsNP Without ESS 64 29 35 Roxithromycin 150 Placebo 12


Videler 2011 Mixed (wNP) Without ESS 60 29 31 Azithromycin 500/7* Placebo 12
Zeng 2011 CRSsNP Without ESS 43 22 21 Clarithromycin 250 INCS 12
Jiang 2012 CRSsNP Without ESS 53 27 26 Erythromycin 500 Herb 8
Peric 2014 CRSwNP ESS:preoperative 80 40 40 Clarithromycin 500 No macrolide 8
Korkmaz 2014 CRSwNP Without ESS 44 22 22 Clarithromycin 250† No macrolide 8
Varvyanskaya 2014 CRSwNP ESS:postoperative 66 44 22 Clarithromycin 250 No macrolide 24
Amali 2015 Mixed (sNP) ESS:postoperative 66 22 44 Azithromycin 250 Placebo 12
Haxel 2015 Mixed (wNP) ESS:postoperative 58 29 29 Erythromycin 250 Placebo 12
Deng 2018 Mixed (wNP) Without ESS 74 38 36 Clarithromycin 250 No macrolide 12

*Study group received azithromycin 500 mg/d for 3 days during the first week followed by 500 mg/wk for 11 weeks.

Study group received clarithromycin 1,000 mg/d during the first 2 weeks, followed by 250 mg/d for 6 weeks.
CRS = chronic rhinosinusitis; CRSsNP = chronic rhinosinusitis without nasal polyps; CRSwNP = chronic rhinosinusitis with nasal polyps; ESS = endoscopic
sinus surgery; Mixed (sNP) = mixed population with predominant without polyps; Mixed (wNP) = mixed population with predominant with polyps; INCS = Intranasal
corticosteroids.

LDMs therapy to a standard treatment of intranasal steroid (SMD = 0.15, 95% CI: −0.25 to 0.54).16,18 Heterogeneity
spray.21 One trial was excluded from quantitative synthesis was substantial for the SNOT (I2 = 88%), symptom score
because the LDMs therapy was compared to herbal medicine, (I2 = 85%), endoscopy score (I2 = 98%). There was no het-
which was neither a placebo nor a standard treatment.17 erogeneity (I2 = 0%) for CT score.

Outcomes Comparison: LDMs Versus Standard Treatment


Eight trials assessed the SNOT. 8,9,12,15–18,20
Seven trials There was only one RCT in this comparison.21 The
assessed symptom scores,8,9,12,16,19–21 seven trials assessed results showed no difference between LDMs and intrana-
endoscopy,8,12,16,17,19–21 two trials assessed radiological sal steroid spray in the improvement of symptom score
scores,16,18 nine trials reported gastrointestinal and car- (MD = 0.04, 95% CI: −0.56 to 0.64) and endoscopy score
diac adverse effects,8,9,12,15–20 and four trials reported data (MD = −0.49, 95% CI: −0.10 to 0.12). The SNOT and CT
after the end of the treatment.8,9,12,19 score were not assessed.21

Prognostic Factor: CRS Subtype


Overall Effects of LDMs Therapy When subgroup analysis by CRS subtype was per-
The data of the findings from the meta-analyses and formed, the effects favored LDMs over placebo in the
subgroup analysis are displayed in Table II. improvement in the SNOT in patients with CRSsNP
(SMD = −0.64, 95% CI: −1.01 to −0.27), but not in patients
with CRSwNP (SMD = 0.18, 95% CI: −0.19 to 0.55). The sub-
Comparison: LDMs Versus Placebo group difference was statistically significant (P = .009); the
The meta-analysis revealed no difference between data are displayed in Figure 2. Likewise, the effects favored
the LDMs and placebo in the improvement in 1) the the LDMs over placebo in the improvement in symptom
SNOT (SMD = −0.23, 95% CI: −0.69 to 0.24),8,9,12,15 2) score in patients with CRSsNP (MD = −0.89, 95% CI: −1.41
symptom score (SMD = −0.29, 95% CI: −1.46 to 0.89),8,12 to −0.37), but not in patients with CRSwNP (SMD = 0.31,
and endoscopy score (SMD = −0.35, 95% CI: −0.71 to 95% CI: −0.21 to 0.83). The subgroup difference was statisti-
0.00).8,12 There was no trial assessing the improvement cally significant (P = .001). There was no difference between
in CT score. Heterogeneity was substantial for the SNOT the two subgroups (P = .64) in endoscopy score.
(I2 = 68%), symptom score (I2 = 90%). There was no het-
erogeneity (I2 = 0%) for the endoscopy score.
Prognostic Factor: Serum IgE Level
Comparison: LDMs Plus Standard Treatment Two RCTs measured serum IgE level at enrollment.8,12
Versus Standard Treatment The serum IgE level prognostic factor could not be assessed
The cumulative meta-analysis revealed no difference because both trials did not report data separately between
between the LDMs plus standard treatment and standard patients with low and high serum IgE level.
treatment in the improvement in 1) the SNOT (SMD = −0.52,
95% CI: −1.57 to 0.53),16,18,20 2) symptom score (SMD = −0.63, Prognostic Factor: Concurrent ESS
95% CI: −1.42 to 0.16),16,19,20 3) endoscopy score (SMD = Compared to placebo, LDMs brought greater symptom
−1.85, 95% CI: −5.59 to 1.88),16,19,20 and 4) CT score improvement when given to patients without ESS (MD =

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


4
TABLE II.
Summary of Findings From Meta-analyses and Subgroup Analyses.
Outcomes
Prognostic Factors Comparison QOL Score Symptom Score Endoscopy Score CT Score

1. CRS subtype LDMs vs. placebo Favor CRSsNP, Favor CRSsNP, N/D N/A
SMD −0.64 SMD −0.89
1.1 CRSwNP
(−1.01, −0.27) (−1.41, −0.37)
1.2 CRSsNP
LDMs plus Standard treatment N/D N/D N/D N/D
vs. standard treatment
LDMs vs. standard treatment N/A N/A N/A N/A
2. Serum IgE LDMs vs. placebo N/A N/A N/A N/A
2.1 low serum IgE LDMs plus standard treatment N/A N/A N/A N/A
2.2 high serum IgE vs. standard treatment
LDMs vs. standard treatment N/A N/A N/A N/A
3. Concurrent ESS LDMs vs. placebo N/D Favor LDMs N/D N/A
without ESS,
3.1 LDMs w/o ESS
SMD −0.89
3.2 LDMs with ESS
(−1.41, −0.37)
LDMs plus standard treatment Favor LDMs with N/D Favor LDMs With N/D
vs. standard treatment ESS, SMD ESS, SMD
−1.68 (−2.40, −3.79 (−4.85,
−0.95) −2.73)
LDMs vs. standard treatment N/D N/D N/D N/D
4. Dose LDMs vs. placebo Favor half dose, Favor half dose, N/D N/A
SMD −0.64 SMD −0.89
4.1 Half dose
(−1.01, −0.27) (−1.41, −0.37)
4.2 Less than half dose
LDMs plus standard treatment N/A N/D N/D N/A
vs. standard treatment
LDMs vs. Standard treatment N/A N/A N/A N/A
5. Lactone ring LDMs vs. placebo N/D N/A N/A N/A
5.1 14-membered ring LDMs plus standard treatment N/A N/A N/A N/A
5.2 15-membered ring vs. standard treatment
LDMs vs. standard treatment N/A N/A N/A N/A
6. Duration LDMs vs. placebo N/A N/A N/A N/A
6.1 > 12 weeks LDMs plus standard treatment Favor >12 weeks, Favor >12 weeks, Favor >12 weeks, N/D
6.2 12 weeks vs. standard treatment SMD −1.68 SMD −1.65 SMD −3.79
6.3 < 12 weeks (−2.40, −0.95) (−2.37, −0.93) (−4.85, −2.73)
LDMs vs. standard treatment N/A N/A N/A N/A

CRS = chronic rhinosinusitis; CRSsNP = chronic rhinosinusitis without nasal polyps; CRSwNP = chronic rhinosinusitis with nasal polyps; ESS = endo-
scopic sinus surgery; LDMs = low-dose macrolides; N/A = not applicable; N/D = no difference between subgroup; QOL = quality of life.

−0.89, 95% CI: −1.41 to −0.37) over patients with ESS 95% CI: −0.46 to 0.51). The subgroup difference was statisti-
(MD = 0.31, 95% CI: −0.21 to 0.83). The subgroup difference cally significant (P < .001). The improvement was not differ-
was statistically significant (P = .001). Improvements were ent between patients with and without ESS in symptom score
similar between patients with and without ESS in the (the patients without ESS, MD = −0.10, 95% CI: −0.59 to
SNOT improvement (patients without ESS, SMD = −0.20, 0.38), and the patients with ESS (SMD = −0.94, 95% CI:
95% CI: −1.03 to 0.63) and the patients with ESS (SMD = −2.27 to 0.39; P = .25).
−0.26, 95% CI: −1.04 to 0.53; P = .76) and endoscopy score
(patients without ESS, MD = −0.27, 95% CI: −0.77 to 0.22),
and patients with ESS (MD = −0.45, 95% CI: −0.97 to Prognostic Factor: Dose of Macrolides
0.08; P = .64). When subgroup analysis by dose of macrolides was
When LDMs plus standard treatment was compared performed, the effects favored the patients receiving a
with standard treatment, meta-analyses gave results in con- half dose of macrolides (SMD = −0.64, 95% CI: −1.01 to
trast to the comparison of LDMs versus placebo. LDMs −0.27), over the patients receiving less than a half dose of
brought greater SNOT and endoscopy score improvements macrolides (SMD = 0.18, 95% CI: −0.19 to 0.55; P = .002).
when given to patients with ESS. The improvement in the The data are displayed in Figure 3. Likewise, the effects
SNOT favored patients with ESS (MD = −1.68, 95% CI: −2.40 in symptom score improvement favored patients receiving
to −0.95) but not the patients without ESS (SMD = −0.01, a half dose of macrolides (MD = −0.89, 95% CI: −1.41 to
95% CI: −0.63 to 0.61). The subgroup difference was statisti- −0.37), over the patients receiving less than a half-dose of
cally significant (P < .001). The improvement in endoscopy macrolides (MD = 0.31, 95% CI: −0.21 to 0.83; P = .001).
score favored the patients with ESS (SMD = −3.79, 95% CI: There was no difference between the two subgroups
−4.85 to −2.73), but not the patients without ESS (MD = 0.02, (P = .64) in endoscopy score.

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


5
Fig. 2. Improvement in the SNOT at the end of treatment when low-dose macrolides therapy was compared with placebo and subgroup analy-
sis by CRS subtype. CI = confidence interval; CRSsNP = chronic rhinosinusitis without polyps; CRSwNP = chronic rhinosinusitis with polyps;
df = degrees of freedom; IV = inverse variance; LDM = low-dose macrolides; Random = random effects; SD = standard deviation; SNOT =
Sino-Nasal Outcome Test; Std. mean difference = standardized mean difference. [Color figure can be viewed in the online issue, which is
available at www.laryngoscope.com.]

Prognostic Factor: Membered Lactone Ring LDMs (SMD = −0.21, 95% CI: −1.09 to 0.66; P = .96. The
of LDMs data are displayed in Figure 4.
When subgroup analysis by membered lactone ring of
LDMs was performed, the improvements in the SNOT were
similar between the patients receiving a 14-membered lac- Prognostic Factor: Duration of Treatment
tone ring of LDMs (SMD = −0.24, 95% CI: −0.98 to 0.50), When subgroup analysis by duration of LDMs treat-
and the patients receiving a 15-membered lactone ring of ment was performed, the effects in SNOT improvement

Fig. 3. Improvement in the SNOT at the end of treatment when low-dose macrolides therapy was compared with placebo and subgroup analy-
sis by dosage. CI = confidence interval; df = degrees of freedom; IV = inverse variance; LDM = low-dose macrolides; Random = random
effects; SD = standard deviation; SNOT = Sino-Nasal Outcome Test; Std. mean difference = standardized mean difference. [Color figure can
be viewed in the online issue, which is available at www.laryngoscope.com.]

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


6
Fig. 4. Improvement in the SNOT at the end of treatment when LDM therapy was compared with placebo and subgroup analysis by membered
ring of macrolides. CI = confidence interval; df = degrees of freedom; Fixed = fixed effects; IV = inverse variance; LDM = low-dose macro-
lides; SD = standard deviation; SNOT = Sino-Nasal Outcome Test; Std. mean difference = standardized mean difference. [Color figure can be
viewed in the online issue, which is available at www.laryngoscope.com.]

favored patients receiving 24-week LDMs (SMD = –1.68, score8,12 (SMD = −0.17, 95% CI: −0.53 to 0.18), and no dif-
95% CI: −2.40 to −0.95), over patients receiving 12-week ference between the LDMs plus standard treatment and
LDMs (MD = −0.28, 95% CI: −0.77 to 0.21) and 8-week standard treatment in the symptom score19 (MD = −0.06,
LDMs (MD = 0.36, 95% CI: −0.33 to 1.04; P = .002). Like- 95% CI: −0.49 to 0.16). At the time point of 48 weeks, Peric
wise, the effects favored patients receiving 24-week LDMs et al.19 reported no difference between the LDMs plus stan-
in symptom improvement (MD = −1.65, 95% CI: −2.37 to dard treatment and standard treatment in the symptom
−0.93), over patients receiving 12-week LDMs (MD = −0.10, score (MD = −0.17, 95% CI: −0.61 to 0.27).
95% CI: −0.59 to 0.38) and 8-week LDMs (MD = −0.29, 95%
CI: −0.73 to 0.15; P = .001), and the effects favored patients
receiving 24-week LDMs in endoscopy score (MD = −3.79,
Risk of Bias of Included Studies
95% CI: −4.85 to −2.73), over patients receiving 12-week
The included studies had substantial selection bias
LDMs (MD = 0.02, 95% CI: −0.46 to 0.51; P < .001). The
for random sequence generation (60% low risk) and allo-
effects at 8 weeks in one RCT were not estimable. The
cation concealment (50% low risk). They had modest risks
improvement in CT score were similar between patients
in detection bias (70% low risk), attrition bias (80% low
receiving 12-week LDMs (MD = 0.08, 95% CI: –0.40 to 0.56),
risk), and reporting bias (80% low risk).
and patients receiving 8-week LDMs (MD = 0.28, 95% CI:
−0.40 to 0.96; P = .64).

Sensitivity Analysis
Adverse Effects The sensitivity analysis was performed by excluding
There were nine studies that reported gastrointesti- studies with multiple (more than one) high risks of bias from
nal and cardiac adverse effects.8,9,12,15–20 LDMs produced the meta-analysis. There were two RCTs excluded.19,20 Both
greater gastrointestinal adverse effects (5%) when com- RCTs compared LDMs plus standard treatment versus
pared to other treatments (1.05%) (risk ratio: 3.52; 95% CI: standard treatment. The results revealed two significant
1.29 to 9.60). There was no cardiac adverse effect reported prognostic factors in that the LDMs were effective in SNOT
in any patients. The data are displayed in Table III. improvement, which were CRSsNP and a half dose of LDMs,
and three prognostic factors in symptom improvement,
which were CRSsNP and a half dose of LDMs and LDMs
therapy without ESS.
Durability of Outcomes Improvement
There were four studies that reported data after the
end of the treatment at the time point of 24 weeks.8,9,12,19
The meta-analysis revealed no difference between the LDMs Meta-regression Analysis
and placebo in the improvement in 1) the SNOT8,9 Meta-regression analysis was not performed due to
(SMD = −0.28, 95% CI: −0.64 to 0.09) and 2) endoscopy the limited number of included studies.

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


7
TABLE III.
Summary of Gastrointestinal and Cardiac Adverse Effects of LDMs Versus Other Treatments.
Gastrointestinal Adverse Effects Cardiac Adverse Effects
LDMs, No. Other Treatments, No. LDMs, No. Other Treatments, No.
First Author Year Event Total Event Total Event Total Event Total

Wallwork 2006 1 29 1 35 0 29 0 35
Videler 2011 2 29 2 31 0 29 0 31
Jiang 2012 0 27 0 26 0 27 0 26
Peric 2014 2 40 0 40 0 40 0 40
Korkmaz 2014 1 22 0 22 0 22 0 22
Varvyanskaya 2014 1 44 0 22 0 44 0 22
Amali 2015 0 22 0 44 0 22 0 44
Haxel 2015 6 29 0 29 0 29 0 29
Deng 2018 1 38 0 36 0 38 0 36
14 280 3 285 0 285 0 285

LDMs = low-dose macrolides.

DISCUSSION patients receiving roxithromycin: low (<200 IU/mL) and


When compared to the controls, the overall effects of high serum IgE. Only the patients with low serum IgE
LDMs for treating CRS did not favor LDMs in the showed improvement in the SNOT after treatment. It is
improvement of any outcomes. To date, there have been worth noticing that the low IgE level defined by this study
four meta-analyses assessing the effects of LDMs for treat- (<200 IU/mL) is greater than the general cut point of
ing CRS. Pynnonen et al.22 pooled data from two RCTs <100 IU/mL. Haxel et al.12 performed subgroup analysis
and showed no benefit of LDMs over placebo at the end of and reported no difference in all outcomes between the
treatment. Head et al.4 extracted data from three RCTs. patients with low (<100 IU/mL) and high serum IgE. More-
The meta-analysis included one RCT for each comparison, over, a recent study by Maniakas et al.26 gave azithromycin
and each outcome showed no benefit of the LDMs therapy. to patients who did not respond to postoperative budesonide
Lasso et al.23 pooled data from four RCTs and showed no irrigation, and found that the macrolide responders had
difference between the effects of LDMs and placebo. The higher mean serum IgE level (208 IU/mL) than the nonre-
recent meta-analysis by Shen et al.24 included both ran- sponders (72 IU/mL). Recently, the local IgE production
domized and nonrandomized controlled trials. Forest plots within the nose and paranasal sinuses in patients with CRS
from the RCTs did not show the benefit of LDMs therapy. has been reported.27,28 Thus, the low serum IgE level may
Although the beneficial effects of LDMs therapy were not be a good predictor to identify macrolide responders.
not evident by meta-analyses, substantial heterogeneity The included studies used various treatment regimens.
was shown. When six predictive factors were assessed by Subgroup analyses in this study showed no difference
our study, subgroup analyses revealed that LDMs were between the 14-membered ring and the 15-membered ring
effective in a specific patient population or optimal treat- LDMs. The anti-inflammatory effects of LDMs that interfere
ment regimens. The CRS subtype and serum IgE were with the cytokine production and inflammatory cell metabo-
assessed in this study based on the mechanism of antineu- lism were revealed in the 14- and 15-membered ring but not
trophilic action and the suppression of the production of IL-8 16-membered ring macrolides.29,30 The hydrophobic nature
and TNF-α of the LDMs therapy. The findings from sub- of the 14- and 15-membered lactone rings alters the biophysi-
group analyses showed that the LDMs therapy had benefi- cal properties of the cell membrane of the effector inflamma-
cial effects in the improvement of the SNOT and symptom tory cells. It interferes with the regulation of intracellular
scores over placebo only in the patients with CRSsNP. These metabolic and transcriptional pathways involved in the
could be explained by the immunopathogenesis of CRSsNP inflammatory cascade.31,32 These immunomodulatory effects
driven by type 1/type 17 cytokines and the inflammatory of macrolides appear to be independent of the antimicrobial
pattern of neutrophilic/noneosinophilic inflammation. On properties, and the dosages of LDMs are much lower than
the other hand, the CRSwNP associates with type 2 cyto- the minimum inhibitory concentration. However, there is no
kines and high tissue eosinophilia.25 Thus, its immuno- clear evidence that shows the optimal dose of LDMs. The
pathogenesis may not respond to the immunomodulation dosages of LDMs used by the included studies varied from
pathway of macrolides. Serum IgE could not be assessed by a half dose to a very low dose. When subgroup analyses
our meta-analyses because the included studies did not were performed, the beneficial effects of LDMs therapy over
report data separately between the patients with low and placebo in the improvement of the SNOT and symptoms
high serum IgE. The serum IgE level is acknowledged as a were shown in the subgroup using half-dose macrolides. This
seromarker for type 2 inflammation, and low serum IgE is in line with in vitro studies showing that the reduction of
level has been recommended by the EPOS 2012 for identify- proinflammatory cytokine production is dose dependent.
ing macrolide responders.1 However, clinical studies showed Kohyama et al.30 and Wallwork et al.33 assessed eosinophils
controversies. Wallwork et al.8 studied two subgroups of and nasal mucosal tissue cultured, respectively, and revealed

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


8
the concentration-dependent reduction in IL-8. Duration of Favorable outcomes may be achieved in both patients receiv-
treatment is the other controversial issue. The included stud- ing macrolides with and without ESS.
ies gave LDMs therapy for various durations of 8, 12, and
24 weeks. The findings from subgroup analyses revealed that
the 24-week duration of LDMs therapy had greater benefits Acknowledgments
than other durations. The treatment of CRS aims to effec- The author thank Likhit Khattiyawittayakun for his tech-
tively control chronic inflammatory conditions of the parana- nical advice.
sal sinuses. Neither corticosteroids nor LDMs therapy aims
to cure the underlying etiologies. Thus, the duration of LDMs
BIBLIOGRAPHY
therapy should not be limited to 12 weeks. Longer duration
1. Fokkens WJ, Lund VJ, Mullol J, et al. European position paper on rhinosi-
of LDMs therapy may achieve better long-term disease nusitis and nasal polyps 2012. Rhinol Suppl 2012;23:1-298.
control by the long-term anti-inflammatory and immuno- 2. Devillier P, Roche N, Faisy C. Clinical pharmacokinetics and pharmacody-
namics of desloratadine, fexofenadine and levocetirizine: a comparative
modulatory effects. review. Clin Pharmacokinet 2008;47:217–230.
Subgroup analyses by concurrent ESS in this study 3. Tomassen P, Vandeplas G, Van Zele T, et al. Inflammatory endotypes of
chronic rhinosinusitis based on cluster analysis of biomarkers. J Allergy
showed mixed results. Patients receiving LDMs without Clin Immunol 2016;137:1449–1456.e1444.
ESS had greater improvement in symptoms than placebo. 4. Head K, Chong LY, Piromchai P, et al. Systemic and topical antibiotics for
chronic rhinosinusitis. Cochrane Database Syst Rev 2016;4:CD011994.
However, the patients receiving LDMs with concurrent 5. Cervin A, Wallwork B. Efficacy and safety of long-term antibiotics (macro-
ESS had greater improvement in the SNOT and endos- lides) for the treatment of chronic rhinosinusitis. Curr Allergy Asthma
copy score. After removing the low-quality studies to per- Rep 2014;14:416.
6. Oakley GM, Harvey RJ, Lund VJ. The role of macrolides in chronic rhinosi-
form sensitivity analysis, the patients without ESS had a nusitis (CRSsNP and CRSwNP). Curr Allergy Asthma Rep 2017;17:30.
greater symptom score improvement. However, the effect 7. Shimizu T, Suzaki H. Past, present and future of macrolide therapy for
chronic rhinosinusitis in Japan. Auris Nasus Larynx 2016;43:131–136.
was too small (0.89) to be clinically meaningful. On the 8. Wallwork B, Coman W, Mackay-Sim A, Greiff L, Cervin A. A double-blind,
other hand, the effects of the SNOT improvement (1.68) randomized, placebo-controlled trial of macrolide in the treatment of
chronic rhinosinusitis. Laryngoscope 2006;116:189–193.
and endoscopy score improvement (3.79) found in patients 9. Videler WJ, Badia L, Harvey RJ, et al. Lack of efficacy of long-term, low-
with concurrent ESS were larger. Overall, the findings dose azithromycin in chronic rhinosinusitis: a randomized controlled trial.
Allergy 2011;66:1457–1468.
suggested the modest beneficial effect of LDMs therapy 10. Orlandi RR, Kingdom TT, Hwang PH, et al. International consensus state-
without ESS was inferior to corticosteroids and ESS. ment on allergy and rhinology: rhinosinusitis. Int Forum Allergy Rhinol
2016;6(suppl 1):S22–S209.
Thus, the LDMs therapy should not be the first-line treat- 11. Rudmik L, Soler ZM. Medical therapies for adult chronic sinusitis: a sys-
ment for patients with CRS. Among the studies of LDMs tematic review. JAMA 2015;314:926–939.
12. Haxel BR, Clemens M, Karaiskaki N, Dippold U, Kettern L, Mann WJ. Con-
with concurrent ESS, Peric et al.19 did not find any bene- trolled trial for long-term low-dose erythromycin after sinus surgery for
fit of preoperative LDMs. The LDMs therapy after ESS chronic rhinosinusitis. Laryngoscope 2015;125:1048–1055.
13. Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interven-
for long-term control of chronic inflammatory conditions tions. Version 5.1.0. Available at: www.cochrane-handbook.org. Updated
of paranasal sinuses, which cannot be managed by sur- March 2011. Accessed 12 April 2018.
gery, is clinically meaningful and more practical. 14. Review Manager (RevMan) [computer program]. Version 5.1.6. Copenhagen,
Denmark: The Nordic Cochrane Centre; 2011.
To the best of our knowledge, this study is the first 15. Amali A, Saedi B, Rahavi-Ezabadi S, Ghazavi H, Hassanpoor N. Long-term
meta-analysis assessing prognostic factors that predict postoperative azithromycin in patients with chronic rhinosinusitis: a ran-
domized clinical trial. Am J Rhinol Allergy 2015;29:421–424.
the success of LDMs therapy in patients with CRS. The 16. Deng J, Chen F, Lai Y, et al. Lack of additional effects of long-term, low-
findings suggested that LDMs therapy provided clinical dose clarithromycin combined treatment compared with topical steroids
alone for chronic rhinosinusitis in China: a randomized, controlled trial.
effectiveness to patients with CRSsNP. LDMs therapy Int Forum Allergy Rhinol 2018;8:8–14.
should be an option for some groups of patients. When 17. Jiang RS, Wu SH, Tsai CC, Li YH, Liang KL. Efficacy of Chinese herbal
medicine compared with a macrolide in the treatment of chronic rhinosi-
LDMs are considered, a half-dose of LDMs for 24 weeks nusitis without nasal polyps. Am J Rhinol Allergy 2012;26:293–297.
duration is recommended. 18. Korkmaz H, Ocal B, Tatar EC, et al. Biofilms in chronic rhinosinusitis with polyps:
is eradication possible? Eur Arch Otorhinolaryngol. 2014;271:2695–2702.
The limitation of this study is that the 10 included 19. Peric A, Baletic N, Milojevic M, et al. Effects of preoperative clarithromycin
studies in this meta-analysis had multiple comparisons administration in patients with nasal polyposis. West Indian Med J 2014;
63:721–727.
with several treatment outcomes. When subgroup analyses 20. Varvyanskaya A, Lopatin A. Efficacy of long-term low-dose macrolide ther-
were performed, the number of patients in each subgroup apy in preventing early recurrence of nasal polyps after endoscopic sinus
surgery. Int Forum Allergy Rhinol 2014;4:533–541.
may not have had enough power to see a statistically sig- 21. Zeng M, Long XB, Cui YH, Liu Z. Comparison of efficacy of mometasone furo-
nificant difference. The heterogeneity was substantial in ate versus clarithromycin in the treatment of chronic rhinosinusitis without
some meta-analyses. Bias among the included studies was nasal polyps in Chinese adults. Am J Rhinol Allergy 2011;25:e203–e207.
22. Pynnonen MA, Venkatraman G, Davis GE. Macrolide therapy for chronic rhi-
demonstrable. nosinusitis: a meta-analysis. Otolaryngol Head Neck Surg 2013;148:366–373.
23. Lasso A, Masoudian P, Quinn JG, et al. Long-term low-dose macrolides for
chronic rhinosinusitis in adults—a systematic review of the literature.
Clin Otolaryngol 2017;42:637–650.
CONCLUSION 24. Shen S, Lou H, Wang C, Zhang L. Macrolide antibiotics in the treatment of
chronic rhinosinusitis: evidence from a meta-analysis. J Thorac Dis 2018;
Although overall beneficial effects were not demon- 10:5913–5923.
strated, LDMs with appropriate treatment regimens may 25. Snidvongs K, Lam M, Sacks R, et al. Structured histopathology profiling of chronic
rhinosinusitis in routine practice. Int Forum Allergy Rhinol 2012;2:376–385.
provide clinical benefits in disease-specific quality of life, 26. Maniakas A, Desrosiers M. Azithromycin add-on therapy in high-risk
symptoms, endoscopy, and radiology to a specific patient pop- postendoscopic sinus surgery patients failing corticosteroid irrigations: a
clinical practice audit. Am J Rhinol Allergy 2014;28:151–155.
ulation. The findings from meta-analyses and subgroup ana- 27. Baba S, Kondo K, Toma-Hirano M, et al. Local increase in IgE and class
lyses suggested that the LDMs should be clinically effective switch recombination to IgE in nasal polyps in chronic rhinosinusitis. Clin
Exp Allergy 2014;44:701–712.
in patients with CRSsNP. When LDMs are administered, a 28. Cao PP, Zhang YN, Liao B, et al. Increased local IgE production induced by
half-dose of macrolides for 24 weeks duration is suggested. common aeroallergens and phenotypic alteration of mast cells in Chinese

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


9
eosinophilic, but not non-eosinophilic, chronic rhinosinusitis with nasal 31. Garey KW, Alwani A, Danziger LH, Rubinstein I. Tissue reparative effects
polyps. Clin Exp Allergy 2014;44:690–700. of macrolide antibiotics in chronic inflammatory sinopulmonary diseases.
29. Kadota J, Mukae H, Ishii H, et al. Long-term efficacy and safety of clari- Chest 2003;123:261–265.
thromycin treatment in patients with diffuse panbronchiolitis. Respir Med 32. Hatipoglu U, Rubinstein I. Low-dose, long-term macrolide therapy in
2003;97:844–850. asthma: an overview. Clin Mol Allergy 2004;2:4.
30. Kohyama T, Takizawa H, Kawasaki S, Akiyama N, Sato M, Ito K. Fourteen- 33. Wallwork B, Coman W, Feron F, Mackay-Sim A, Cervin A. Clarithromycin
member macrolides inhibit interleukin-8 release by human eosinophils and prednisolone inhibit cytokine production in chronic rhinosinusitis.
from atopic donors. Antimicrob Agents Chemother 1999;43:907–911. Laryngoscope 2002;112:1827–1830.

Laryngoscope 00: 2019 Seresirikachorn et al.: Predicting Success of Low-Dose Macrolides


10

You might also like