American Journal of Respiratory and Critical Care Medicine

Rationale: Chronic obstructive pulmonary disease (COPD) exacerbations are a major cause of morbidity and mortality, and preventing them is a key treatment target. Long-term macrolide treatment is effective at reducing exacerbations, but there is a paucity of evidence for other antibiotic classes.

Objectives: To assess whether 12-month use of doxycycline reduces the exacerbation rate in people with COPD.

Methods: People with moderate to very severe COPD and an exacerbation history were recruited from three UK centers and randomized to 12 months of doxycycline 100 mg once daily or placebo. The primary study outcome was the exacerbation rate per person-year.

Results: A total of 222 people were randomized. Baseline mean FEV1 was 1.35 L (SD, 0.35 L), 52.5% predicted (SD, 15.9% predicted). The median number of treated exacerbations in the year before the study was 2 (SD, 1–4). A total of 71% of patients reported two or more exacerbations, and 81% were already prescribed inhaled corticosteroids at baseline. The COPD exacerbation rate did not differ between the groups (doxycycline/placebo rate ratio [RR], 0.86; 95% confidence interval [CI], 0.67–1.10; P = 0.23). No difference was seen if only treated exacerbations or hospitalizations were considered. In preplanned subgroup analysis, doxycycline appeared to better reduce the exacerbation rate among people with severe COPD (RR, 0.36; 95% CI, 0.15–0.85; P = 0.019) and in those with an eosinophil count <300 cells/μl (RR, 0.50; 95% CI, 0.29–0.84; P = 0.01). Health status measured by St. George’s Respiratory Questionnaire was 5.2 points worse in the doxycycline group at 12 months (P < 0.007).

Conclusions: Doxycycline did not significantly reduce the exacerbation rate, over 12 months, in participants with COPD who exacerbated regularly, but it may have benefitted those with more severe COPD or blood eosinophil counts <300 cells/μl.

Clinical trial registered with (NCT02305940).

Scientific Knowledge on the Subject

Current clinical guidelines recommend considering prophylactic azithromycin therapy to reduce the frequency of chronic obstructive pulmonary disease (COPD) exacerbations. Azithromycin is sometimes contraindicated and in practice some patients receive prophylactic doxycycline despite sparse evidence to support this choice. Further, few studies consider interactions between peripheral eosinophil count at baseline (non-exacerbating state) and the effect of antibiotic prophylaxis on exacerbation frequency.

What This Study Adds to the Field

Overall, this 12-month, multicentre, double-blind, randomized placebo-controlled trial found that prophylactic doxycycline did not significantly reduce the exacerbation rate in participants with moderate to very-severe COPD who exacerbated regularly. However, prophylactic doxycycline may prove more effective for those with lower baseline eosinophil rate or severe disease.

Acute exacerbations of chronic obstructive pulmonary disease (COPD) are common (1) and drive healthcare use, hospitalization, disease progression (2, 3), and mortality (4). Preventing exacerbations is a key approach toward improving quality of life and reducing healthcare use (5). Preventative, long-term antibiotic treatment has been shown to prevent exacerbations among those with COPD (6). Most trials have studied macrolides (predominantly azithromycin) (713), although one studied moxifloxacin (14), and one four-armed trial studied azithromycin, doxycycline, and moxifloxacin (15). Guidelines support the use of long-term macrolides to prevent exacerbations (5), but azithromycin is contraindicated in some patients, and the effectiveness of other antibiotics remains uncertain. Furthermore, except for current smokers, for whom azithromycin appears to be ineffective, it is unclear which subgroups of patients benefit most from long-term antibiotic therapy (7).

Although the effectiveness of long-term nonmacrolide antibiotics remains uncertain, we found that doxycycline was the second most common long-term antibiotic therapy used by patients with COPD in the United Kingdom (16). Doxycycline possesses both bacteriostatic antibacterial properties and complex antiinflammatory actions (17). Long-term doxycycline is already used for several nonrespiratory conditions, can reduce airway bacteria in patients with COPD, and compares favorably with azithromycin and moxifloxacin regarding adverse events, bacterial resistance, and adherence profiles (15). However, the beneficial effect of macrolides may relate in part to their immunomodulatory effects, a feature not shared by many other antibiotic classes (18).

Studies assessing the efficacy of nonmacrolide antibiotics, in particular doxycycline, are needed to help clinicians decide which long-term antibiotics to prescribe and identify those patients who might benefit from such therapy (6, 19). In this study, we assessed whether 12-month use of doxycycline reduced the COPD exacerbation rate and whether there were different responses among particular patient subgroups.

Study Design

We enrolled outpatients attending three UK centers: the Royal Brompton Hospital (London), University Hospital Aintree (Liverpool), and St. George’s Hospital (London) from both primary and secondary care.

This was a double-blind, randomized placebo-controlled trial of 12 months of doxycycline 100 mg daily conducted between July 15, 2014, and July 12, 2017. The primary study outcome was COPD exacerbation rate per person-year. Secondary outcomes included time to first COPD exacerbation; annual rate of corticosteroid- and/or antibiotic-treated exacerbations; number and duration of COPD-related hospital admissions; and change in spirometry, respiratory health status (measured by St. George’s Respiratory Questionnaire [SGRQ]), and C-reactive protein during the study period. The relationship between the primary outcome and important covariates was explored.

Research ethics approval was granted by the Hampstead Research Ethics Committee, and all patients provided their written informed consent.

Inclusion and Exclusion Criteria

Patients enrolled were aged 45 years or older, had Global Initiative for Chronic Obstructive Lung Disease grade 2–4 (moderate to very severe) COPD (post-bronchodilator ratio of FEV1 to FVC of <0.7 and a post-bronchodilator FEV1 < 80% predicted), reported at least a 10 pack-year history of smoking, and had received treatment with antibiotics and/or corticosteroids for at least one COPD exacerbation in the preceding 12 months. All patients were enrolled when clinically stable, being well for at least 4 weeks after their last COPD exacerbation. Exclusion criteria are listed in the online supplement.


Patient medical history, concomitant medications, and oxygen saturations were recorded; post-bronchodilator spirometry measurements were made according to American Thoracic Society/European Respiratory Society criteria (20); and an ECG was performed. Blood samples were collected to record full blood count and C-reactive protein and to assess safety (liver function, renal function, and coagulation). Patients were trained to complete daily symptom diary cards, to record any change in treatment (antibiotics and/or steroids), and to record adverse events, as previously reported (21). Patients also completed the SGRQ and the COPD Assessment Tool (CAT). Chronic bronchitis was defined as a productive cough lasting at least 3 months per year for at least 2 consecutive years.


Patients were randomized 1:1 to treatment or control (placebo) groups using a computer-generated permuted block system of variable sizes. Randomization was stratified by site and smoking status (current or ex-smoker) to ensure group balance within each site based on smoking status. Over-encapsulated placebo and doxycycline (100 mg) tablets were visually identical. Unblinding for analysis was undertaken after data entry was completed and locked.


Patients were reviewed face-to-face at 3, 6, 9, and 12 months. At all visits, diary cards were reviewed, and any exacerbations, treatment for exacerbations, and/or adverse events were assessed and recorded. Study medication compliance was assessed from pill counts of remaining medication at each visit. Participants were contacted by telephone 2 weeks after enrollment and within 1 month after study completion to further record adverse events.

At the final visit (12 mo), concomitant medications were recorded, spirometry was repeated, and the SGRQ questionnaires were completed. Blood sampling was repeated at each visit. Patients were asked to attend follow-up visits, even if they prematurely discontinued the study treatment.

Definition of Exacerbations and Treatment

The onset of an exacerbation was defined from diary card data as the development or worsening of at least two symptoms (including at least one major symptom) that lasted for at least 2 consecutive days (11). Major symptoms included dyspnea, increased sputum volume, and increased sputum purulence. Minor symptoms included sore throat, cold (nasal discharge and/or nasal congestion), fever (>37.5°C), increased cough, or increased wheeze or chest tightness. An exacerbation was defined as lasting from the day of onset until the last day before 2 consecutive symptom-free days.

In the absence of a completed diary card, where available, patient-reported healthcare use was also used to define exacerbation occurrence. In such cases, the exacerbation onset was defined as the first day of treatment with antibiotics and/or steroids or the first day of hospitalization. As in previous studies, exacerbation identification from diary cards was adjudicated in a blinded manner by an author not involved in data collection during the study (11).

Statistical Analysis

We estimated that enrollment of 156 patients would provide 90% power to detect a significant difference in exacerbation rates between the doxycycline and placebo groups using Poisson regression. This assumed an exacerbation rate ratio (RR) (erythromycin to placebo) of 0.648 from Seemungal and colleagues (11), a mean exacerbation rate of 2.26 (G. C. Donaldson, unpublished results), and an overdispersion of 1.29 from Hurst and colleagues (22). Assuming a 30% dropout or nonadherence rate, we planned to recruit a sample of 222 individuals.

Differences in the exacerbation rate per person-year between the doxycycline and placebo groups were analyzed using negative binomial regression taking account of time periods when patients are exacerbating and, hence, cannot develop an exacerbation, and adjusting for both the number of exacerbations in the previous year and by site (23). The mean exacerbation rate in each group was estimated by dividing the total number of exacerbations in each group by the follow-up time of each group. Supporting analyses included adjustments for baseline covariates (FEV1, SGRQ score, and sex) and then adjusting further for additional covariates that proved unbalanced at baseline (24).

Regarding the primary outcome, differences in therapy impact were assessed according to belonging to the following subgroups: age, sex, baseline smoking status (current or ex-smoker), COPD severity (FEV1 percent predicted < 30% vs. 30–50%, <30% vs. >50%), baseline blood eosinophil count (<300 vs. ⩾300 cells/μl), and exacerbation rate within the previous year (0–1 vs. ⩾2).

Secondary endpoints were analyzed as sensitivity analyses. Three analyses were undertaken for each endpoint. First, models were adjusted for the baseline value of the studied variable alongside the study center. Second, models were further adjusted for baseline values imbalanced at baseline (exacerbation rate in preceding year, FEV1 percent predicted, SGRQ score, and sex). Third, models were further adjusted for enrollment FEV1/FVC, FEV1, and oxygen saturation.

Data are presented as mean and SD or 95% confidence interval (CI), as appropriate. All tests were two-sided, and significance was set at P < 0.05. In the event of the primary outcome being nonsignificant, all other P values reported were considered to be of only nominal significance.

Study Population

Study flow and completion is shown in Figure 1, with 355 people screened, 222 randomized, and 183 completing the study (18% dropout). There were no differences in the withdrawal rates between the two study arms (online supplement). From the 222 randomized patients, 107 were enrolled at the Royal Brompton Hospital (London), 111 at Aintree Hospital (Liverpool), and 4 at St. George’s Hospital (London).

The baseline characteristics of the randomized patients are shown in Table 1. Mean age was 67 years, 57% were male, and 34% were current smokers. The doxycycline groups had greater tobacco exposure (50 vs. 41 pack-years). The FEV1/FVC ratio was slightly lower in the doxycycline group (FEV1/FVC, 0.43 vs. 0.47). Mean FEV1 and FEV1 percent predicted had lower values in the doxycycline relative to the placebo group (FEV1, 1.25 L vs. 1.44 L; FEV1 percent predicted, 50.6% vs. 54.4%, respectively). The number (percentage) of subjects with moderate, severe, and very severe COPD was 121 (55%), 83 (37%), and 18 (8%), respectively.

Table 1. Demographic Features of the Study Population Reported for All Participants (and by Treatment Arm) at Enrollment

Number of subjects222112110
Age, yr, mean (SD)66.7 (7.6)66.1 (7.6)67.3 (7.7)
Males, n (%)126 (57)62 (55)64 (58)
Weight, kg, median (IQR)74.4 (62.3–88.0)74.3 (62.8–92.9)74.5 (61.3–84.1)
BMI, kg/m2, median (IQR)26.3 (23.2–30.9)26.5 (23.3–32.0)26.2 (23.2–30.4)
Current smokers, n (%)75 (34)38 (34)36 (34)
Pack-years, median (IQR)45 (30–56)41 (30–50)50 (35–58)
Previous regular drug smoker, n (%)24 (11)11 (10)13 (12)
FEV1, L, mean (SD)1.35 (0.53)1.44 (0.56)1.25 (0.48)
FEV1 < 30%, n (%)18 (8)7 (6)11 (10)
FEV1 30–49.9%, n (%)83 (37)38 (34)45 (41)
FEV1 50–80%, n (%)121 (55)67 (60)54 (49)
FEV1 percent predicted, mean (SD)52.5 (15.9)54.4 (15.8)50.6 (15.8)
FEV1/FVC, mean (SD)0.45 (0.13)0.47 (0.12)0.43 (0.13)
Oxygen saturation, median (IQR)95 (94–97)96 (94–97)95 (93–97)
Chronic bronchitis, n (%)137 (62)67 (60)70 (64)
Number of treated exacerbations in prior 12 mo, median (IQR)2 (1–4)3 (1–4)2 (1–4)
⩾2 exacerbations in prior 12 mo, n (%)158 (71)82 (73)76 (69)
CAT score, mean (SD)21.1 (8)21.4 (8.4)20.8 (7.7)
SGRQ total, mean (SD)51.7 (19.4)51.3 (19.5)52.2 (19.4)
SGRQ symptoms, mean (SD)56.7 (23.4)56.5 (24.8)56.9 (21.9)
SGRQ activity, mean (SD)69.1 (22.1)68.5 (21.4)69.7 (22.8)
SGRQ impacts, mean (SD)40.1 (20.7)39.7 (20.7)40.5 (20.7)
Prescribed ICS, n (%)180 (81)89 (79)91 (83)
Prescribed triple therapy (LAMA, LABA with ICS), n (%)144 (65)74 (66)70 (64)
Prescribed long-term oral prednisolone, n (%)4 (2)04 (4)
CRP, mg/L, mean (SD)4 (6)3 (5)5 (6)
Eosinophil count, 109/L, median (IQR)0.20 (0.10–0.30)0.20 (0.10–0.30)0.20 (0.10–0.30)
Eosinophil count ⩾0.3 × 109/L, n (%)59 (27)30 (27)29 (26)

Definition of abbreviations: BMI = body mass index; CAT = COPD assessment test; COPD = chronic obstructive pulmonary disease; CRP = C-reactive protein; ICS = inhaled corticosteroid; IQR = interquartile range; LABA = long-acting beta-agonist; LAMA = long-acting muscarinic-antagonist; SGRQ = St. George’s Respiratory Questionnaire.

Across the year preceding enrollment, subjects recalled suffering a median of two exacerbations (interquartile range [IQR], 1–4), with 158 (71%) individuals recalling at least two exacerbations. Subjects reported a high symptom burden and poor health status (mean CAT score, 21.1; mean SGRQ score, 51.7). At enrollment, 180 (81%) subjects were already prescribed inhaled corticosteroids, and 144 (65%) were using triple inhaled therapy. Baseline prescriptions appeared similar across the doxycycline and placebo groups, but four subjects, all randomized to receive doxycycline, were already prescribed long-term oral corticosteroids at enrollment. At enrollment, the blood eosinophil count was ⩾300 cells/μl among 59 (27%) subjects and <300 cells/μl among 163 (73%) subjects and was similar across treatment arms (blood eosinophil count was ⩾300 cells/μl in 26% of subjects in the doxycycline group and 27% in the placebo group).

Overall, FEV1 and FEV1 percent predicted values were lower among subjects enrolled in Liverpool relative to London (mean FEV1, 1.21 [0.50] L vs. 1.48 [0.52] L; mean FEV1 percent predicted, 50.5 [15.9]% vs. 54.5 [15.7]%).

Exacerbation Rate

During the study, subjects experienced a calculated median of 3.2 (IQR, 1.0–5.4) exacerbations per person-year, with a median of 1.4 (IQR, 0–3.4) treated exacerbations per person-year.

There was no statistically significant difference in the arithmetically calculated overall COPD exacerbation rate between the doxycycline and placebo groups: median, 2.28 (IQR, 1.03–4.68) versus 3.35 (IQR, 1.03–5.68) events per person year, respectively, resulting in an unadjusted negative binomial RR of 0.86 (95% CI, 0.67–1.10; P = 0.23). Adjustment for baseline smoking status, exacerbation rate during the previous year (number of exacerbations or frequent [⩾2] versus not frequent [<2] exacerbation status), sex, and SGRQ did not change this conclusion; the adjusted RR was 0.85 (95% CI, 0.67–1.07; P = 0.16). Additional analysis adjusting for study site, lung function, and oxygen saturations at baseline also did not significantly alter these results.

The rate of treated exacerbations did not differ significantly between the doxycycline and placebo groups: median (IQR) 1.27 (0–3.21) versus 2.01 (0–3.91) events per person-year, respectively, giving a RR of 0.87 (95% CI, 0.65–1.16; P = 0.34) by negative binomial analysis, a finding unchanged after adjustment for the covariates listed above. Doxycycline did not alter the number of hospitalizations (26 in the doxycycline vs. 21 in the placebo group) or time spent in the hospital. Figure 2 shows the Kaplan-Meier plot of the time to first exacerbation, with groups stratified according to treatment and enrollment site. There was no significant difference in the time to first exacerbation between the groups.

In preplanned subgroup analyses, we examined the effect of doxycycline according to COPD severity (moderate or severe vs. very severe), eosinophil count (<300 vs. ⩾300 cells/μl), smoking status (current smokers vs. ex-smokers), age (>67 vs. ⩽67 yr), and sex (male vs. female).

Significant interactions were detected between baseline COPD severity and treatment arm, suggesting doxycycline had a greater impact on the exacerbation rate among those with severe COPD. Among those enrolled with severe COPD (FEV1, 30–50% predicted), doxycycline reduced the exacerbation rate by a factor of 0.36 (95% CI, 0.15–0.85; P = 0.019) relative to those with very severe COPD (FEV1 < 30% predicted). No significantly greater doxycycline effect was detected among those with moderate COPD (FEV1, 50–80% predicted), the exacerbation rate factor being 0.51 (95% CI, 0.22–1.17; P = 0.11). These data are shown in Figure 3.

A significant interaction was also detected between enrollment blood eosinophil count and treatment. Doxycycline had a bigger impact on those with eosinophil counts <300 cells/μl relative to those with ⩾300 eosinophils/μl. Among those with a blood eosinophil count at enrollment of <300 cells/μl, doxycycline reduced the exacerbation rate by a factor of 0.50 (95% CI, 0.29–0.84; P = 0.01) relative to those with eosinophil counts ⩾300 cells/μl. These data are shown in Figure 3.

With respect to the primary endpoint, no interaction was detected between the treatment arm and smoking status, age, or sex (Figure 3).

Other Secondary Outcomes

There was no effect of doxycycline on FEV1 at the end of the study. Health status, measured by SGRQ, was worse at the end of the study among those taking doxycycline relative to placebo by 5.2 (95% CI, 1.44–9.00) points (P = 0.007), when adjusted for enrollment site, baseline SGRQ, and time on study. Health status worsened slightly in those treated with doxycycline, improved in those treated with placebo, and was not affected by blood eosinophil count.

Adverse Events

There was no substantial difference in the overall number of adverse events, with 270 in the placebo subjects and 250 in the doxycycline subjects (Table 2). Gastrointestinal disorders were more common with doxycycline (39 vs. 25) because of the more frequent occurrence of dyspepsia, nausea, and vomiting, but this was not significant statistically. The occurrence of diarrhea was the same in both groups. There was no difference in hepatic, dermatological, or cardiac adverse events, but nervous system adverse events were more common in placebo subjects (21 vs. 9), as were infections (55 vs. 28).

Table 2. Adverse Events and Median Time on Study According to Study Arm

 Placebo GroupDoxycycline Group
Total adverse events270250
Serious adverse events4845
Nonserious adverse events222205
 Cardiac disorders1113
 Ear and labyrinth disorders05
 Eye disorders52
 Gastrointestinal disorders2539
  Dyspepsia and gastritis512
  Nausea and vomiting16
 Hepatobiliary disorders33
  Candida infection32
  Dental infection73
  Ear infection54
  Urine infection42
  Bacterial infection, not specified1911
 Injury or poisoning disorder1624
 Metabolic or nutrition disorder1311
 Musculoskeletal disorder3329
 Nervous system disorder219
 Renal or urinary disorders47
 Respiratory disorders1110
 Skin or cutaneous disorders1410
  Rash, not specified42
  Eczema or dermatitis14
 Surgical or medical procedures2520
 Vascular disorders23
Time on study of all participants randomized into each study arm, d, median (IQR)400 (21)398 (20)

Specific adverse event categories are included if at least 1% of subjects experienced that type of adverse event. Specific events are also included (shown in italics) if ⩾5% of subjects had that adverse event.

In this randomized, double-blind, placebo-controlled trial enrolling patients with at least moderate COPD who experienced one or more exacerbations during the previous year, prophylactic doxycycline did not significantly change the exacerbation rate or lung function and was associated with a worse health status at 1 year compared with placebo therapy. However, the therapeutic impact of doxycycline varied according to patient characteristics at study entry. Compared with those with very severe COPD, patients with severe COPD appeared to benefit from antibiotic treatment, whereas those with moderate COPD did not benefit. In addition, doxycycline reduced exacerbation rates among patients with baseline eosinophil counts <300 cells/μl when compared with patients with eosinophil counts of ⩾300 cells/μl. Although not helpful for every patient, these findings suggest doxycycline may reduce the COPD exacerbation rate among specific patient subgroups.

This is the first randomized controlled study to investigate the effect of prophylaxis with doxycycline on future COPD exacerbations. A recent Cochrane meta-analysis of 14 completed trials aiming to reduce exacerbations of COPD using antibiotic prophylaxis found long-term antibiotics reduced exacerbations by a factor of 0.57 (95% CI, 0.42–0.78; P < 0.001) (6). These trials mostly studied how macrolides reduce exacerbation rates, and international guidelines advocate using azithromycin prophylaxis for this purpose (5, 25). Our study design and sample compare favorably with these preceding studies. Although smaller than the largest macrolide study (7), our study is larger, similar in design, and contains a similar study population to the positive macrolide study performed by Seemungal and colleagues (11). With a study sample of just 109 patients, Seemungal and colleagues were able to show that erythromycin significantly reduced the exacerbation rate, with a RR relative to placebo of 0.65. Although Seemungal and colleagues focused on moderate or severe exacerbations, our negative finding was unaltered after restricting our analysis to treated exacerbations only. Significant macrolide effects have been found in even smaller study samples enriched for more frequent exacerbators (26); however, the proportion of our subjects who experienced at least three exacerbations per year already appears comparable to Seemungal and colleagues (47% vs. 35%) (11). Therefore, our findings suggest that long-term doxycycline is broadly less effective than macrolides at reducing COPD exacerbation frequency.

Our negative primary outcome is important because, despite limited evidence, long-term doxycycline has been more commonly prescribed than azithromycin to patients with COPD within the United Kingdom (16). Although our headline findings may deter this generic approach, this may also obscure a useful role for doxycycline within certain subsets of patients. Age, smoking status, and sex did not influence the response to doxycycline, but we found that doxycycline reduced exacerbations among those with severe COPD better than among those with moderate or very severe COPD. Weaker exacerbation reduction among those with worse spirometry measurements has been observed previously in bronchodilator studies (27, 28), perhaps reflecting differences in exacerbation pathophysiology accompanying increasing disease severity.

Our data also suggest that the effectiveness of doxycycline in reducing exacerbations was influenced by the patient’s baseline eosinophil count. Although not considered in previous comparable COPD prophylactic antibiotic studies (7, 8, 11, 14), a similar interaction has been observed in one 3-month study of azithromycin and trials of long-term antibiotics targeting other inflammatory respiratory diseases. In a post hoc analysis, Vermeersch and colleagues found that long-term azithromycin commenced at the onset of a severe exacerbation reduced treatment failure, particularly among those with blood eosinophil counts <300 cells/μl at the exacerbation onset (29). An elevated blood eosinophil count in chronic rhinosinusitis was found to predict a poor response to long-term macrolide therapy (30), and long-term doxycycline may worsen outcomes among those with coexisting asthma or elevated baseline serum IgE concentrations (31). Furthermore, although tetracyclines offer immunomodulatory benefits to some individuals with asthma, these benefits may be absent among those with coexisting COPD (32).

The mechanism underlying the interaction between doxycycline and baseline eosinophil count deserves further study and may reflect the dual bacteriostatic and antiinflammatory action of tetracyclines (33, 34). Doxycycline’s bacteriostatic properties may prove most beneficial to patients with lower eosinophil counts, perhaps treating an associated heightened susceptibility to infection contributing to exacerbation occurrence (35). However, in other patients, baseline eosinophilic inflammation may be a more prominent driver of exacerbations (35). Current COPD treatment strategies use a blood eosinophil count threshold of 300 cells/μl (36) to help identify steroid responsiveness (37, 38). Plausibly, eosinophil concentrations might also predict the response to other antiinflammatory agents, such as tetracyclines. In vivo, doxycycline exhibits antiinflammatory effects (39), including downregulation of eosinophil degranulation (40) and decreased nitric oxide production (41), which could theoretically impair defense against some infections. Variation in the airway microbiome is also linked to blood eosinophil counts, and Proteobacteria may dominate in those with low eosinophil counts. Suppression of these organisms with doxycycline therapy may be beneficial, but a reduction in the Firmicutes and Streptococcus dominating at higher eosinophil counts might be disadvantageous (42). Future trials will be needed to test such hypotheses.

At enrollment, our patients reported significant impairment of health status, as identified by both the SGRQ and CAT questionnaires. SGRQ did not change in those receiving doxycycline but improved in those randomized to placebo. These differences may be explained by the selective withdrawal of the sicker patients taking placebo, as observed in the trials of inhaled corticosteroids (43). The difference is not explained by blood eosinophil count; health status worsened to a similar degree in those receiving doxycycline and improved to a similar degree in those receiving placebo, regardless of whether the baseline blood eosinophil count was higher or lower. The pattern of adverse events reported with doxycycline was similar to that of the placebo treatment. The reasons behind the apparent failure of doxycycline to improve health status remain unclear.

A major strength of our study is that patients reported a significant number of exacerbations during their observation period, identified using a well-tried methodology that allowed us to compute an event rate in an unbiased way (23). Although not powered to investigate the subset of exacerbations that were treated, this study showed similar trends, albeit statistically nonsignificant, toward doxycycline reducing the exacerbation rate using either definition. As expected, because of their prior history of exacerbations, most of our patients were already using inhaled corticosteroid and bronchodilator therapy, and this precluded an exploration of interactions between these treatments and maintenance antibiotic therapy. Furthermore, we were not able to conduct detailed surveillance of antibiotic resistance in our patients, a concern that may limit the wider use of antibiotic prophylaxis (15).

In conclusion, doxycycline 100 mg daily did not significantly reduce overall exacerbation frequency or improve quality among those with at least moderate COPD. However, the impact of long-term doxycycline varied according to the baseline blood eosinophil count. Future studies are needed to confirm whether the potentially substantial effect of doxycycline on exacerbation numbers is present in patients prospectively identified as having a blood eosinophil count <300 cells/μl.

The authors thank the patients who participated in the study, trial coordinator Dr. Ethaar El-Emir, and the research nurses. They also thank their hospitals for support, the members of the data monitoring and safety committee (Tim Clayton, Nick Hopkinson, Dan Jackson, Patrick Mallia, and Robert Stockley), and John Hurst and Irwin Nazareth, who provided constructive comments on the trial steering committee.

1. Vogelmeier CF, Criner GJ, Martinez FJ, Anzueto A, Barnes PJ, Bourbeau J, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive lung disease 2017 report. GOLD executive summary. Am J Respir Crit Care Med 2017;195:557582.
2. Seemungal TA, Donaldson GC, Paul EA, Bestall JC, Jeffries DJ, Wedzicha JA. Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998;157:14181422.
3. Donaldson GC, Seemungal TA, Bhowmik A, Wedzicha JA. Relationship between exacerbation frequency and lung function decline in chronic obstructive pulmonary disease. Thorax 2002;57:847852.
4. Soler-Cataluña JJ, Martínez-García MA, Román Sánchez P, Salcedo E, Navarro M, Ochando R. Severe acute exacerbations and mortality in patients with chronic obstructive pulmonary disease. Thorax 2005;60:925931.
5. National Clinical Guideline Centre. Chronic obstructive pulmonary disease in over 16s: diagnosis and management. National Institute for Health and Care Excellence; 2018 [accessed Dec 2022]. Available from:
6. Herath SC, Normansell R, Maisey S, Poole P. Prophylactic antibiotic therapy for chronic obstructive pulmonary disease (COPD). Cochrane Database Syst Rev 2018;10:CD009764.
7. Albert RK, Connett J, Bailey WC, Casaburi R, Cooper JA Jr, Criner GJ, et al.; COPD Clinical Research Network. Azithromycin for prevention of exacerbations of COPD. N Engl J Med 2011;365:689698.
8. Simpson JL, Powell H, Baines KJ, Milne D, Coxson HO, Hansbro PM, et al. The effect of azithromycin in adults with stable neutrophilic COPD: a double blind randomised, placebo controlled trial. PLoS One 2014;9:e105609.
9. Wang P, Yang J, Yang Y, Ding Z. Effect of azithromycin in combination with simvastatin in the treatment of chronic obstructive pulmonary disease complicated by pulmonary arterial hypertension. Pak J Med Sci 2017;33:260264.
10. He ZY, Ou LM, Zhang JQ, Bai J, Liu GN, Li MH, et al. Effect of 6 months of erythromycin treatment on inflammatory cells in induced sputum and exacerbations in chronic obstructive pulmonary disease. Respiration 2010;80:445452.
11. Seemungal TA, Wilkinson TM, Hurst JR, Perera WR, Sapsford RJ, Wedzicha JA. Long-term erythromycin therapy is associated with decreased chronic obstructive pulmonary disease exacerbations. Am J Respir Crit Care Med 2008;178:11391147.
12. Suzuki T, Yanai M, Yamaya M, Satoh-Nakagawa T, Sekizawa K, Ishida S, et al. Erythromycin and common cold in COPD. Chest 2001;120:730733.
13. Janjua S, Mathioudakis AG, Fortescue R, Walker RA, Sharif S, Threapleton CJ, et al. Prophylactic antibiotics for adults with chronic obstructive pulmonary disease: a network meta-analysis. Cochrane Database Syst Rev 2021;1:CD013198.
14. Sethi S, Jones PW, Theron MS, Miravitlles M, Rubinstein E, Wedzicha JA, et al.; PULSE Study group. Pulsed moxifloxacin for the prevention of exacerbations of chronic obstructive pulmonary disease: a randomized controlled trial. Respir Res 2010;11:10.
15. Brill SE, Law M, El-Emir E, Allinson JP, James P, Maddox V, et al. Effects of different antibiotic classes on airway bacteria in stable COPD using culture and molecular techniques: a randomised controlled trial. Thorax 2015;70:930938.
16. James GD, Petersen I, Nazareth I, Wedzicha JA, Donaldson GC. Use of long-term antibiotic treatment in COPD patients in the UK: a retrospective cohort study. Prim Care Respir J 2013;22:271277.
17. Sapadin AN, Fleischmajer R. Tetracyclines: nonantibiotic properties and their clinical implications. J Am Acad Dermatol 2006;54:258265.
18. Martinez FJ, Curtis JL, Albert R. Role of macrolide therapy in chronic obstructive pulmonary disease. Int J Chron Obstruct Pulmon Dis 2008;3:331350.
19. Allinson JP, Donaldson GC. Long-term antibiotic therapy reduces exacerbation frequency in patients with COPD but it remains unclear which patients to target. Evid Based Med 2014;19:99.
20. Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A, et al.; ATS/ERS Task Force. Standardisation of spirometry. Eur Respir J 2005;26:319338.
21. Seemungal T, Harper-Owen R, Bhowmik A, Moric I, Sanderson G, Message S, et al. Respiratory viruses, symptoms, and inflammatory markers in acute exacerbations and stable chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;164: 16181623.
22. Hurst JR, Donaldson GC, Quint JK, Goldring JJ, Baghai-Ravary R, Wedzicha JA. Temporal clustering of exacerbations in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2009;179:369374.
23. Law M, Sweeting MJ, Donaldson GC, Wedzicha JA. Misspecification of at-risk periods and distributional assumptions in estimating COPD exacerbation rates: the resultant bias in treatment effect estimation. Pharm Stat 2017;16:201209.
24. Donaldson GC, Müllerova H, Locantore N, Hurst JR, Calverley PM, Vestbo J, et al. Factors associated with change in exacerbation frequency in COPD. Respir Res 2013;14:79.
25. Wedzicha JA, Calverley PMA, Albert RK, Anzueto A, Criner GJ, Hurst JR, et al. Prevention of COPD exacerbations: a European Respiratory Society/American Thoracic Society guideline. Eur Respir J 2017;50:1602265.
26. Uzun S, Djamin RS, Kluytmans JA, Mulder PG, van’t Veer NE, Ermens AA, et al. Azithromycin maintenance treatment in patients with frequent exacerbations of chronic obstructive pulmonary disease (COLUMBUS): a randomised, double-blind, placebo-controlled trial. Lancet Respir Med 2014;2:361368.
27. Wedzicha JA, Banerji D, Chapman KR, Vestbo J, Roche N, Ayers RT, et al.; FLAME Investigators. Indacaterol-glycopyrronium versus salmeterol-fluticasone for COPD. N Engl J Med 2016;374: 22222234.
28. Calverley PMA, Anzueto AR, Carter K, Grönke L, Hallmann C, Jenkins C, et al. Tiotropium and olodaterol in the prevention of chronic obstructive pulmonary disease exacerbations (DYNAGITO): a double-blind, randomised, parallel-group, active-controlled trial. Lancet Respir Med 2018;6:337344.
29. Vermeersch K, Belmans A, Bogaerts K, Gyselinck I, Cardinaels N, Gabrovska M, et al.; BACE trial investigators. Treatment failure and hospital readmissions in severe COPD exacerbations treated with azithromycin versus placebo: a post-hoc analysis of the BACE randomized controlled trial. Respir Res 2019;20:237.
30. Suzuki H, Ikeda K, Honma R, Gotoh S, Oshima T, Furukawa M, et al. Prognostic factors of chronic rhinosinusitis under long-term low-dose macrolide therapy. ORL J Otorhinolaryngol Relat Spec 2000;62:121127.
31. Pinto Bezerra Soter AC, Bezerra TF, Pezato R, Teles Abdo TR, Pilan RM, Pinna FR, et al. Prospective open-label evaluation of long-term low-dose doxycycline for difficult-to-treat chronic rhinosinusitis with nasal polyps. Rhinology 2017;55:175180.
32. Joks R, Durkin HG. Non-antibiotic properties of tetracyclines as anti-allergy and asthma drugs. Pharmacol Res 2011;64:602609.
33. Labro MT. Immunomodulation by antibacterial agents: is it clinically relevant? Drugs 1993;45:319328.
34. Djekic UV, Gaggar A, Weathington NM. Attacking the multi-tiered proteolytic pathology of COPD: new insights from basic and translational studies. Pharmacol Ther 2009;121:132146.
35. Wedzicha JA, Brill SE, Allinson JP, Donaldson GC. Mechanisms and impact of the frequent exacerbator phenotype in chronic obstructive pulmonary disease. BMC Med 2013;11:181.
36. Calverley PMA, Tetzlaff K, Vogelmeier C, Fabbri LM, Magnussen H, Wouters EFM, et al. Eosinophilia, frequent exacerbations, and steroid response in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2017;196:12191221.
37. Singh D, Agusti A, Anzueto A, Barnes PJ, Bourbeau J, Celli BR, et al. Global strategy for the diagnosis, management, and prevention of chronic obstructive lung disease: the GOLD science committee report 2019. Eur Respir J 2019;53:1900164.
38. Bafadhel M, Peterson S, De Blas MA, Calverley PM, Rennard SI, Richter K, et al. Predictors of exacerbation risk and response to budesonide in patients with chronic obstructive pulmonary disease: a post-hoc analysis of three randomised trials. Lancet Respir Med 2018;6:117126.
39. Watz H, Tetzlaff K, Wouters EF, Kirsten A, Magnussen H, Rodriguez-Roisin R, et al. Blood eosinophil count and exacerbations in severe chronic obstructive pulmonary disease after withdrawal of inhaled corticosteroids: a post-hoc analysis of the WISDOM trial. Lancet Respir Med 2016;4:390398.
40. Van Zele T, Gevaert P, Holtappels G, Beule A, Wormald PJ, Mayr S, et al. Oral steroids and doxycycline: two different approaches to treat nasal polyps. J Allergy Clin Immunol 2010;125:10691076.e4.
41. Hoyt JC, Ballering J, Numanami H, Hayden JM, Robbins RA. Doxycycline modulates nitric oxide production in murine lung epithelial cells. J Immunol 2006;176:567572.
42. Dicker AJ, Huang JTJ, Lonergan M, Keir HR, Fong CJ, Tan B, et al. The sputum microbiome, airway inflammation, and mortality in chronic obstructive pulmonary disease. J Allergy Clin Immunol 2021;147:158167.
43. Calverley PM, Spencer S, Willits L, Burge PS, Jones PW; IOSLDE Study Group. Withdrawal from treatment as an outcome in the ISOLDE study of COPD. Chest 2003;124:13501356.
Correspondence and requests for reprints should be addressed to James P. Allinson, M.D., Ph.D., COPD Research Group, Airways Disease Section, National Heart and Lung Institute, Imperial College London, Guy Scadding Building, Dovehouse Street, London SW3 6LY, UK. E-mail: .

*These authors contributed equally to this work.


Supported by National Institute of Health Research (UK) program grant award RP-PG-0109-10056.

Author Contributions: J.P.A., B.H.V., S.E.B., P.P.W., P.M.A.C., G.C.D., and J.A.W. contributed to the study design, protocol, and study materials. J.P.A., B.H.V., S.E.B., L.J.F., L.A.-M., G.B., and P.P.W. contributed to patient recruitment and collection of study data at participating centers. M.L. and G.B. performed the statistical analysis. J.P.A. wrote the first draft of the manuscript. All authors contributed to interpretation of the data and revision of the manuscript.

This article has a related editorial.

This article has an online supplement, which is accessible from this issue’s table of contents at

Originally Published in Press as DOI: 10.1164/rccm.202212-2287OC on July 14, 2023

Author disclosures are available with the text of this article at

Comments Post a Comment

New User Registration

Not Yet Registered?
Benefits of Registration Include:
 •  A Unique User Profile that will allow you to manage your current subscriptions (including online access)
 •  The ability to create favorites lists down to the article level
 •  The ability to customize email alerts to receive specific notifications about the topics you care most about and special offers
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record