Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Chlamydia pneumoniae and chronic asthma: Updated systematic review and meta-analysis of population attributable risk



Chlamydia pneumoniae (Cp) is an obligate intracellular human respiratory pathogen producing persisting lung infection with a plausible link to asthma pathogenesis. The population attributable risk of potentially treatable Cp infection in asthma has not been reported.


The author searched from 2000 to 2020 inclusive for previously un-reviewed and new cross sectional and prospective controlled studies of Cp biomarkers and chronic asthma in both children and adults. Qualitative descriptive results and quantitative estimates of population attributable risk for selected biomarkers (specific IgG, IgA and IgE) are presented.


No large, long-term prospective population-based studies of Cp infection and asthma were identified. About half of case-control studies reported one or more significant associations of Cp biomarkers and chronic asthma. Heterogeneity of results by age group (pediatric v adult asthma), severity category (severe/uncontrolled, moderate/partly controlled, mild/controlled) and antibody isotype (specific IgG, IgA, IgE) were suggested by the qualitative results and confirmed by meta-analyses. The population attributable risks for Cp-specific IgG and IgA were nul in children and were 6% (95% confidence interval 2%-10%, p = 0.002) and 13% (9%-18%, p<0.00001) respectively in adults. In contrast to the nul or small population attributable risks for Cp-specific IgG and IgA, the population attributable risk for C. pneumoniae-specific IgE (children and adults combined) was 47% (39%-55%, p<0.00001). In the subset of studies that reported on asthma severity categories, Cp biomarkers were positively and significantly (P<0.00001) associated with asthma severity.


C. pneumoniae-specific IgE is strongly associated with asthma and asthma severity, suggesting a possible mechanism linking chronic Cp infection with asthma in a subset of individuals with asthma. Infection biomarkers should be included in future macrolide treatment trials for severe and uncontrolled asthma.


Macrolides are now included in the Global Initiative for Asthma (GINA) [1], European Respiratory Society/American Thoracic Society (ERS/ATS) [2], and British Thoracic Society (BTS) [3] guidelines as treatment options for severe asthma. Macrolide mechanism(s) of action in asthma are currently unknown and could include immunomodulatory, anti-viral, or anti-microbial effects [4]. Information regarding the potential quantitative contribution of candidate organisms to asthma pathogenesis may aid in prioritizing research efforts into macrolide mechanisms of action and possibly help guide clinician prescribing of macrolides for severe asthma. Among the candidate target organisms is Chlamydia pneumoniae (Cp). Cp is an obligate intracellular human pathogen that was first described as a new species causing acute human respiratory infections (primarily bronchitis and pneumonia) in 1986 [5]. The first report associating possible Cp persistent infection or re-infection with asthma was published in 1991 [6]. Two subsequent reviews on associations of Cp and asthma, published in 1999 [7] and 2005 [8], reported on a growing body of evidence linking Cp biomarkers and chronic airway obstructive illnesses (asthma, chronic bronchitis and chronic obstructive pulmonary disease–COPD). Both reviews reported positive associations of infection biomarkers with asthma and highlighted the urgent need for further research to inform treatment guidelines. Neither of the reviews attempted to quantify the potential contribution of infection to asthma population burden. The purpose of the current systematic review and meta-analysis is to (1) tabulate observational studies of Cp biomarkers and chronic asthma in both children and adults that were published during and after 2005—the date of the most recent review [8]–and (2) calculate estimates for the population attributable risk (PAR) of selected Cp biomarkers in chronic asthma.


The overall research question is What proportion of chronic stable asthma is potentially attributable to or influenced by chronic Chlamydia pneumoniae infection? Or In individuals diagnosed with chronic stable asthma, compared to non-asthma controls, do selected Cp infection biomarkers demonstrate an increased prevalence? Because direct microbiologic organism detection of deep lung infection is currently not feasible in population studies, this updated systematic review and meta-analysis focused on peripheral blood serologic biomarkers as potential surrogates for infection.

Search strategy and selection criteria

For the updated systematic review, the author searched Pub Med, Scopus, CINAHL and The Cochrane Library using the search term “asthma and (Chlamydophila pneumoniae or Chlamydia pneumoniae)” limited to the years 2000 to October 2020. (see supplementary file S1 Appendix. Search algorithms for updated systematic review). The years 2000 to 2005 were included in case some studies published in those years had been overlooked by the 2005 review. The search returned 332 unique articles that were screened for inclusion as either case-control or prospective cohort studies of chronic asthma. To avoid confounding by acute infections in acute exacerbations, studies that reported only patients in exacerbation were excluded. Additionally, studies containing a mix of exacerbated and stable asthma patients, that did not report on chronic stable asthma separately, were excluded. Foreign language publications were included if they contained an English language abstract and sufficient biomarker data on cases and controls. A study was included in this updated review if it (1) reported on one or more biomarkers of Cp infection, (2) studied a population of subjects with clinician diagnosed chronic stable asthma, (3) included a non-asthmatic control group and (4) presented sufficient information to calculate biomarker prevalence in the chronic stable asthma case group and control groups separately. An additional criterion for inclusion as a prospective study was a follow up duration of at least five years. The author did not re-tabulate studies that had been reviewed in detail earlier (i.e., were contained in Table I of reference number 8).

Using these criteria 20 publications were included in this updated systematic review (Fig 1). Cross sectional studies were abstracted for study location (country), case and control definitions, microbiological method(s), main finding(s) and additional comments. Prospective studies were abstracted for location, study design, population studied, microbiological method(s), outcome definition(s), main finding(s) and additional comments. Microbiologic tests included culture, polymerase chain reaction (PCR) testing and/or various serologic tests. Selected methodologic issues for case-control studies are discussed in the text and presented systematically in supplementary file S2 Appendix for the studies included in the meta-analysis. This study did not involve human subjects and human subjects committee approval was not sought.

Fig 1. Pub Med search results (asthma and (Chlamydia pneumoniae or Chlamydophila pneumoniae)) limits 1/1/2000 to 10/2020.

Data analysis

Meta-analyses of population attributable risks (PAR) were estimated using Review Manager (RevMan) {Computer program} Version 5.4, The Cochrane Collaboration, 2020. Individual meta-analyses for Cp-specific IgG, IgA and IgE are reported. Appropriate data for meta-analysis were available from a total of 25 studies: 14 tabulated in this updated review and 11 tabulated in the 2005 review [8] (Fig 1). PAR was defined as the absolute difference between prevalence of biomarker positivity in the asthma case group and the non-asthma control group. For example, if 50% of asthma cases and 40% of controls are biomarker positive, the difference (50% minus 40% = 10%) is the population attributable risk. The PAR is an estimate of what percent of an outcome could possibly be prevented if a risk factor were to be removed from a population [9]. The risk difference between cases and controls was analyzed using the Mantel-Haenzel Fixed Effect method. Heterogeneity was assessed using the heterogeneity Chi-square test and the I2 statistic. Results are presented with 95% confidence intervals. Cp-specific IgG, IgA and IgE on peripheral blood were chosen for meta-analysis because they were by far the most commonly reported biomarkers. Peripheral blood is also the most accessible and reliable source for biomarker data in clinical practice and research. Cp IgM was not analyzed because it is primarily a marker for acute, not chronic, infection. Cp IgG and IgA were analyzed individually for child- and adult asthma separately. Due to the limited number of studies, Cp IgE was analyzed for both age groups combined. A subset of studies reported on asthma severity. Due to the limited number of such studies, a meta-analysis combining biomarkers and asthma age groups was performed to estimate and compare the PARs for severity subgroups. Study definitions for severity varied, e.g., some studies used a mild-moderate-severe nosology and others used a controlled-partly controlled-uncontrolled nosology. Therefore, severity was concatenated into mild/controlled, moderate/partly controlled and severe/uncontrolled categories. Biomarker classification (positive or negative), asthma case definitions and severity categories were recorded based on the criteria of the publication.

Quality assessment of individual studies

The Cochrane Collaboration provides a formal risk of bias (ROB) assessment tool in RevMan that is appropriate for randomized trials and includes domains such as random sequence generation, allocation sequence concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data and selective outcome reporting. These domains do not apply to the assessment of case-control studies. For the individual case-control studies included in this meta-analysis four domains were distilled and tailored from the NIH Quality Assessment of Case-Control Studies checklist [10]: (1) case selection bias (Are asthma patients with particular characteristics, e.g. smoking and lung co-morbidities, systematically excluded or included?), (2) case diagnostic certainty (Is asthma diagnosis supported by objective evidence of reversible airway obstruction or airway hyperreactivity?), (3) control selection bias (Are controls representative of the general population?) and (4) biomarker validity (How well are the included biomarker tests validated?). Regarding publication bias, the number of studies were generally insufficient to justify using funnel plots to assess for publication bias according to Cochrane Collaboration guidance [11]. A single analysis (Fig 3A) was examined via a funnel plot because it was the only analytic group containing more than 10 studies. Sensitivity analysis was performed for selected analytic groups (Figs 2A, 3A and 5) to investigate subgroup heterogeneity.

Fig 2. Meta-analysis of population attributable risk (C. pneumoniae biomarkers and asthma) in pediatric asthma.

A: C. pneumoniae-specific IgG. B: C. pneumoniae-specific IgA. Population attributable risk varies from 0 (0%: contributes to none of the disease) to 1 (100%: contributes to all of the disease).

Fig 3. Meta-analysis of population attributable risk (C. pneumoniae biomarkers and asthma) in adult asthma.

A: C. pneumoniae-specific IgG. B: C. pneumoniae-specific IgA. Population attributable risk varies from 0 (0%: contributes to none of the disease) to 1 (100%: contributes to all of the disease).


Updated systematic review

Ten pediatric asthma studies from 7 countries met inclusion criteria (Table 1) [1221]. Nine studies reported peripheral blood biomarkers (serology in 8, culture in 1) and 1 reported on nasal and/or induced sputum specimens. Five of 10 reported one or more positive associations with asthma (Table 1). Two positive studies included severe asthma patients undergoing bronchoscopy for clinical indications [12,20] and a third positive study reported that biomarker positivity tended to be higher in more severe asthma [17]. Another positive study reported that Cp biomarker positivity was associated with non-atopic asthma [14]. The fifth positive study compared atopic asthma with non-atopic controls, raising the possibility of confounding by atopic status [13]. Of the five studies reporting no significant biomarker association with chronic asthma, one reported a positive association of Cp IgM and acute exacerbations (P = 0.03) [16], and another found a positive association of Cp IgA and asthma in the subgroup with inhalant allergy (P = 0.002) but not in the subgroup with food allergy [18]. A third negative study reported a significant (P<0.001) association of Cp IgG with increased attacks but did not provide primary data [21]. A fourth negative study used as controls a group of patients some of whom had sinusitis, otitis and pneumonia [15] and the fifth negative study did not include severe asthma [19].

Table 1. Pediatric case-control studies of Chlamydia pneumoniae biomarkers in asthma.

Ten adult asthma studies from 8 countries met inclusion criteria (Table 2) [2231]. Eight studies reported peripheral blood biomarkers (all serology) and 2 reported on nasal and/or induced sputum specimens. Five of 10 studies reported one or more positive associations with asthma (Table 2). Two positive studies reported significant associations of Cp biomarkers and asthma severity [30] or poor control [29] while the three other positive studies did not document the level of asthma severity/control. One negative study found a significant association of Cp PCR positivity and lower FEV1/FVC (P = 0.02) [24]. Another negative study used family members as controls instead of general community members [26]. In two other negative studies asthma was mild [25] or did not require active symptoms [28]. The fifth negative study failed to include a full array of biomarker data [22].

Table 2. Adult case-control studies of Chlamydia pneumoniae biomarkers in asthma.

No prospective studies meeting inclusion criteria were found. One prospective study in adults was derived from an historical cohort organized for other purposes and failed to meet case definition criteria for asthma [32]. Three pediatric studies followed small numbers of children for less than 2 years and failed to meet duration criterion [3335]. Thus, there were no long-term prospective studies designed specifically to investigate Cp infection in asthma.


The meta-analyses included 11 studies published prior to 2005 [3646] that were tabulated in the 2005 review [8] and not re-tabulated, and 14 studies published since 2005 [1521,23,25,26,2831] that are presented in Tables 1 and 2. Fig 2 presents meta-analyses of the PARs of Cp IgG and Cp IgA for pediatric asthma. There were no significant associations with asthma for either antibody in pediatric asthma (i.e., the PAR confidence intervals include 0%). For IgG there appeared to be significant underlying heterogeneity (I2 = 76%, P = 0.0001). The heterogeneity for Cp IgG is completely explained by three outliers: I2 became zero when Wazir 2007 [17], Annagur 2007 [16] and Tutanç 2015 [21] were removed from the meta-analysis. Wazir 2007 [17] contained almost two-thirds of patients with moderate to severe asthma and was the only one of the studies in Fig 2A to specify asthma severity. The contribution of asthma severity category to the heterogeneity is explored further in the analysis depicted in Fig 5.

Fig 3 presents comparable results for adult asthma. Cp IgG was associated with a significant but small PAR (6%, 95% confidence interval 2%-10%, p = 0.002). Cp IgA was associated with a significant somewhat larger PAR (13%, 9%-18%, p<0.00001). Both estimates exhibited a moderate amount of statistical heterogeneity (I2 = 54% and 34%, respectively). Similar to the case for children, the heterogeneity for Cp IgG is completely explained by outliers: I2 became zero when Sirmatel 2003 [44] and Specjalski 2001 [29] were removed from the meta-analysis. In Specjalski 2001 [29] almost three-quarters of the patients had moderate to severe asthma. Sirmatel 2003 [44] did not provide asthma severity data. A funnel plot identifies these two studies as lying outside the funnel (S1 Fig). This asymmetry is not necessarily related to publication bias; rather it may possibly be explained by a greater than average severity, as explored further in analyses associated with Fig 5.

Fig 4 presents results for Cp IgE for combined age groups. In contrast to the IgG and IgA results, the PAR for IgE was large (47%, 39%-55%, p<0.00001) with only moderate heterogeneity (I2 = 47%, P = 0.05) despite combining child and adult asthma.

Fig 4. Meta-analysis of population attributable risk for C. pneumoniae-specific IgE in pediatric and adult asthma combined.

Pediatric asthma: Patel 2012. Adult asthma: Hahn 2020 & Smith-Norowitz 2020. Population attributable risk varies from 0 (0%: contributes to none of the disease) to 1 (100%: contributes to all of the disease).

Fig 5 presents the meta-analytic results for studies that reported on asthma severity categories. This analysis combined asthma age groups and used all available antibody results (IgG, IgA, IgE). There was a strong positive association of Cp biomarker positivity and asthma severity category that was highly significant (P<0.00001). Cp biomarkers were associated with a 5% PAR (0%-10%, p = 0.03) in mild/controlled asthma, a 28% PAR (19%-36%, p<0.00001) in moderate/partly controlled asthma and a 39% PAR (32%-47%, p<0.00001) in severe/uncontrolled asthma. As one might predict when combining age groups and isotypes, within subgroup heterogeneity ranged from moderate (I2 = 46%) for the mild/controlled subgroup to substantial (I2 = 77%) for the severe/uncontrolled subgroup. The exception was the moderate/partly controlled subgroup that exhibited negligible (I2 = 24%) heterogeneity. Importantly, the PARs between subgroups were highly significantly different (p<0.00001), confirming that severity was likely responsible for some of the heterogeneity noted in previous results.

Fig 5. Meta-analysis of population attributable risk for C. pneumoniae-specific IgG, IgA & IgE in pediatric and adult asthma by severity subgroups.

Pediatric asthma: Patel 2012. Adult asthma: All others. Population attributable risk varies from 0 (0%: contributes to none of the disease) to 1 (100%: contributes to all of the disease).

Interestingly, when Cp IgE studies are removed from the meta-analysis (S2 Fig), the heterogeneity for Moderate/Partly Controlled and Severe/Uncontrolled categories disappears (I2 = 0). Nevertheless, the significant associations of severity category and the other Cp biomarkers remain significant (P<0.00001), despite the fact that the PARs diminish from 28% to 25% for Moderate/Partly Controlled and from 39% to 29% for Severe/Uncontrolled asthma. These patterns suggest that the associations are robust across biomarker types and that Cp IgE produces the strongest associations.

Study quality

Quality assessment results for the studies included in the meta-analyses are presented in supplementary file S2 Appendix. Case selection was rated as good in a minority (5 of 12) of adult asthma studies that could be assessed, indicating that external validity was adversely affected by exclusion of patients who smoked and/or had lung comorbidities. An additional 4 adult studies failed to provide enough information to allow a rating, an indication of inadequate reporting quality. The majority of studies (15 of 23) confirmed the asthma diagnosis with objective measures and only 2 studies failed to provide data, indicating that diagnostic certainty was generally good. Most (n = 17) control groups were composed of convenience samples of a healthy (usually patient) population. Only 3 studies reported choosing controls randomly and 5 studies were rated as having made poor choices of controls on the basis of using family members, health care workers, or using inappropriate matching criteria (e.g., intentional mismatches between atopic and non-topic cases and controls). Biomarker validity was rated as Good for 15 studies that employed a version of the microimmunofluorescence (MIF) test or other tests that had been validated against organism detection in either that study or other publications. Nine studies employed an ELISA test of unknown validity and one study did not provide any test information.


This systematic review/meta-analysis focused on chronic asthma, not asthma exacerbations, and was undertaken to update observational studies published since 2005, the date of the last review [8] and, for the first time, to estimate the population attributable risk (PAR) of Cp biomarkers in chronic asthma. Studies of short-term azalide/ketolide treatment for asthma exacerbations have been disappointing [47,48] whereas studies of long-term macrolide/azalide treatment for long-term management of chronic asthma have yielded clinically significant benefits [49], hence the focus on chronic asthma herein. Underlying mechanism(s) of action for macrolide benefits in chronic asthma are unknown and may include pathogen and host-directed anti-inflammatory effects [50]. The meta-analyses found nul to small associations of Cp IgG and IgA with asthma in children and adults, respectively. These results are not surprising given that children often do not mount a detectable antibody response in proven Cp infection [12] and the background prevalence of Cp antibody in the adult population is high [51]. These results suggest that Cp IgG and IgA may not prove clinically useful in predicting which patients will benefit from macrolide treatment. Further analysis found that Cp IgE population attributable risk approached 50% (Fig 4). This result suggests that Cp chronic infection cannot be discounted as a potentially significant contributing factor to asthma, particularly moderate to severe asthma, in both children and adults.

Population attributable risk (PAR) as an outcome was chosen over the odds ratio (OR) because PAR is more clinically interpretable. Patterns and significance of results did not change when OR was substituted as an outcome measure in each meta-analysis (data not shown). Neither Cp IgG nor IgA had a significant association with PAR in child asthma, and had only modest associations (6% and 13%, respectively) in adult asthma. Cp IgG indicates past infection, which is prevalent in adult populations worldwide [51]. Cp IgA is thought to better reflect possible ongoing chronic infection [39]. This study did not assess the value of combined biomarkers (e.g. IgG≥1:xxx and IgA≥1:xx). By contrast with IgG and IgA results, the PAR for Cp IgE was a surprisingly high 47% (95% CI 40% to 55%) with little heterogeneity despite combining asthma age groups. The pathogenicity of antigen-specific IgE in asthma is well established and antigen avoidance and/or removal is the foundation of asthma symptom prevention and treatment. Asthma, whether atopic or non-atopic (i.e., associated or not with skin test positivity against a battery of aeroallergens) is almost always associated with some type of IgE-related reaction [52]. The results of this meta-analysis raise the possibility that some undetected antigen(s) responsible for generating a cryptic IgE response in some asthma patients could be expressed by Chlamydia pneumoniae.

Whether Cp IgE reflects ongoing Cp infection or a “hit and run” phenomenon is not established. Bronchoscopic studies of severely symptomatic children with asthma and/or chronic bronchitis have demonstrated the presence of Cp (by PCR, staining and/or culture) in around half the studied patients [12,20,53]. One of these studies reported that elevation of total IgE was strongly associated with lavage culture positivity of Chlamydia [12]. These observations suggest that Cp IgE should be explored further for its accuracy in predicting the presence of ongoing treatable chronic Cp infection in asthma. There are no comparable bronchoscopic studies of adults with severe asthma. Cp has been detected within lung tissue monocytes by immunostaining in 44% of young adults (mean age 32 years) autopsied after traumatic deaths [54] and Cp PCR positivity in peripheral blood is equally prevalent in healthy blood donors [55,56]. Thus, chronic exposure to Cp antigens appears to be common. Perhaps the presence of Cp IgE in asthma reflects an aberrant host immune response that is causally related to reactive airways diseases. Cp IgM, IgG and IgA are the commercially available biomarkers used to assess acute Cp infection in medical practice. The results of this study suggest that these commercially available tests have limited sensitivity and specificity to identify potentially chronically infected individuals. Cp IgE may be a more promising biomarker to explore in patients with asthma.

Severe asthma is currently a guideline criterion for macrolide treatment [13]. Fig 5 presents evidence that Cp biomarkers, whether IgG, IgA or IgE, are associated with increasing asthma severity in a somewhat linear fashion. This finding has two implications: (1) Cp infection is a potential underlying cause for macrolide benefits in a subset of patients with severe asthma and (2) some proportion of less than severe asthma may also be attributable to infection. It has not been ruled out that macrolides (and perhaps other antibiotics, or combinations of antibiotics) are beneficial for some selected patients with moderate/partly controlled or even mild/controlled, asthma. Based on the data in Fig 5, the estimated numbers needed to treat (NNT) are 20 for mild/controlled asthma, 4 for moderate/partly controlled asthma, and 3 for severe/uncontrolled asthma. These estimates assume that antibiotics are completely effective in removing the risk factor and that the risk factor is causal or a necessary contributing factor. The former assumption is unlikely to be true and the latter is plausible but unproven. The results do support the need however to include Cp biomarkers (especially Cp IgE) in future macrolide treatment trials, and to include less than severe asthma in such trials. Clinicians should keep in mind the possibility that they are treating an infection as they prescribe macrolides for asthma, and consider following a recommendation of the British Thoracic Society guidelines [3] that “ongoing treatment should be guided by clinical response based on specific outcome measures, including exacerbation frequency, symptoms and quality of life assessed at baseline.”


This review and meta-analysis did not address a potential role for Mycoplasma pneumoniae (Mp). Mp is associated with asthma exacerbations, mostly but not exclusively in children [57], and has recently been associated with asthma inception in a population-based study [58]. This review and meta-analysis did not address the considerable amount of in vitro and in vivo experimental data supporting the plausible role for Cp in asthma pathogenesis, extensively reviewed elsewhere [5961]. This study has limitations including issues with diversity of diagnosis, patient selection, and study and biomarker methodology in the source material, all of which may contribute to analytic heterogeneity (i.e., the possibility that an analysis is missing important subgroup differences and/or comparing “apples and oranges”). The main requirement for inclusion in this study was clinician-diagnosed asthma. Some, but not all, studies included objective evidence for reversible airway obstruction. Up to 30% of physician diagnosed asthma patients may not have the diagnosis confirmed after systematic evaluation [62]. Overdiagnosis would tend to decrease contrasts, i.e., bias towards nul findings and is less likely as asthma severity increases [62]. Variations in choice of the control group likely increased heterogeneity and probably also biased towards the nul. For example, Kocabas et al. [26] used relatives of cases as controls, which probably biased results towards the nul because family members are likely to share viruses and bacteria. Another source of bias is the all too common practice in asthma case-control studies to conflate asthma and atopy by comparing, for example, an atopic case group with a non-atopic control group (or vice versa), as was done by Kopriva et al. [13]. Variability in biomarker methodology included differences in laboratory test methods and use of different criteria for positive and negative results. This variability underscores the fact that there are no established serological criteria for chronic infection [63]. All these limitations likely contributed to the heterogeneity of results. Despite many of these limitations biasing towards the nul the results for Cp IgE and severity were quantitatively large. An assessment of the prevalence of these biases can be found in the supplementary S2 Appendix.

Study quality assessment

The Cochrane Collaboration has developed a tool for Risk of Bias (ROB) assessment of randomized trials of medical interventions but has not developed a quality tool for case-control studies. Therefore, the case-control studies included in the meta-analyses reported here were assessed using a tool developed for this report that included four domains relevant to the assessment of case-control studies: case selection bias, case diagnostic certainty, control selection bias and biomarker validity. Case selection bias assessed the extent to which the included cases were representative of typical asthma patients in the population, as a measure of generalizability or external validity. Asthma efficacy trials upon which the guideline treatments are based systematically exclude asthma patients who smoke, or who have ever smoked more than 10 pack-years, or who suffer from lung co-morbidities such as chronic bronchitis and/or emphysema [64,65]. These types of selection bias are critically important to acknowledge because biomarkers of C. pneumoniae infection are associated with smoking, chronic bronchitis and emphysema [7,66,67]. The case selection bias domain qualitatively assessed the degree to which such exclusions were implemented in individual case-control studies: a Good rating was given to studies that included smoking and co-morbidities, a Fair rating was given to studies that included one but not both and a Poor rating was given to studies that excluded both smoking and lung co-morbidities. The case diagnostic certainty domain assessed whether the asthma diagnosis was supported by objective evidence of either reversible airway obstruction measured by pulmonary function testing or bronchial hyperreactivity testing. Clinician diagnosed asthma–the primary inclusion criterion in this report–may be incorrect in up to 30% of diagnosed patients [62]. Such overdiagnosis is more common in less severe asthma and may decrease power to show case-control differences. A Good rating was applied to studies that included systematic objective airway testing and a Fair rating was given to studies without such systematic testing. Control selection bias assessed the extent to which the included controls were representative of typical individuals without asthma in the general population. Choosing randomly selected healthy individuals received a Good rating. Selecting a convenience sample of healthy non-asthma controls was given a Fair rating. Selecting family members, health care workers or subjects with respiratory illnesses as controls received a Poor rating because transmission and/or prevalence of C. pneumoniae infection is enhanced in these settings [59]. Biomarker validation assessed the extent to which the serologic test(s) employed had been validated against human infections documented by organism detection. A Good rating was given, e.g., to the microimmunofluorescence (MIF) test that has been assessed against human population infections confirmed by organism detection [68]. A Fair rating was given to enzyme-linked immunosorbent assays (ELISAs) or other serological tests for which no comparable published validation data could be found. Overall quality assessment results confirm the need for larger, better designed studies into the potential role of chronic C. pneumoniae infection in asthma. Ideally such studies should include direct organism identification from the lung but this ideal appears infeasible currently and thus indirect biomarkers of infection such as serology will be required pending advances in chlamydial sampling and diagnostics for chronic lung infection.

Publication bias is a recognized significant issue for reports of randomized trials of interventions. It is not clear that publication bias affects the reporting of case-control studies to a comparable extent because there are few if any financial disincentives to reporting negative results of case-control studies compared to randomized trials of medical treatments. Except for a single funnel plot associated with Fig 3A, no formal assessment of publication bias was possible for this report due to the limited number of published studies. It is worth noting the paucity of any type of study into this potentially important area of inquiry [69]. In particular, this review highlights the lack of any long-term, population-based prospective observational studies of C. pneumoniae infection in asthma inception or severity.

In summary, this updated systematic review found that biomarkers of Chlamydia pneumoniae infection are associated with chronic asthma in half of observational studies and heterogeneity was suggested based on asthma age group, disease severity category and/or antibody isotype. Meta-analysis results confirmed that age (child asthma v adult asthma), severity (mild/controlled, moderate/partly controlled, severe/uncontrolled) and isotype (IgG, IgA, IgE) were all significantly associated with heterogeneous population attributable risks for asthma. Results also suggest that potential “bacterial allergy” (i.e., presence of Cp-specific IgE) may be a quantitatively important pathogenetic mechanism in asthma, particularly in moderate/partly controlled and severe/uncontrolled asthma. These findings should inform future research into macrolide mechanisms of action in asthma and may inform clinicians’ macrolide treatment plans for selected patients.

Supporting information

S1 Fig. Funnel plot of 14 studies analyzed in Fig 3A.


S2 Fig. Results of sensitivity analysis of severity subgroups (Fig 5) with Cp IgE seroreactivity results removed.


S1 Appendix. Search algorithms for updated systematic review.


S2 Appendix. Quality assessment table for case-control studies included in the meta-analyses.



  1. 1. GINA. Difficult-to-Treat and Severe Asthma in Adolescent and Adult Patients: Diagnosis and Management. 2019;, accessed June 20, 2020.
  2. 2. Holguin F, Cardet JC, Chung KF, Diver S, Ferreira DS, Fitzpatrick A, et al. Management of severe asthma: a European Respiratory Society/American Thoracic Society guideline. Eur Respir J. 2020;55(1). Epub 2019/09/29. pmid:31558662.
  3. 3. Smith D, Du Rand IA, Addy C, Collyns T, Hart S, Mitchelmore P, et al. British Thoracic Society guideline for the use of long-term macrolides in adults with respiratory disease. BMJ Open Respir Res. 2020;7(1). Epub 2020/04/26. pmid:32332022; PubMed Central PMCID: PMC7204798.
  4. 4. Wong EH, Porter JD, Edwards MR, Johnston SL. The role of macrolides in asthma: current evidence and future directions. Lancet Respir Med. 2014;2(8):657–70. pmid:24948430.
  5. 5. Grayston JT, Kuo C-C, Wang S-P, Altman J. A new Chlamydia psittaci strain, TWAR, isolated in acute respiratory tract infections. NEJM. 1986;315(3):161–8. pmid:3724806
  6. 6. Hahn DL, Dodge R, Golubjatnikov R. Association of Chlamydia pneumoniae (strain TWAR) infection with wheezing, asthmatic bronchitis and adult-onset asthma. JAMA. 1991;266:225–30. PubMed Central PMCID: PMC2056624. pmid:2056624
  7. 7. Hahn DL. Chlamydia pneumoniae, asthma and COPD: what is the evidence? Ann Allergy Asthma Immunol. 1999;83:271–92. pmid:10541419
  8. 8. Johnston SL, Martin RJ. Chlamydophila pneumoniae and Mycoplasma pneumoniae. A role in asthma pathogenesis? Am J Respir Crit Care Med. 2005;172:1078–89. pmid:15961690
  9. 9. Uter W, Pfahlberg A. The application of methods to quantify attributable risk in medical practice. Stat Methods Med Res. 2001;10(3):231–7. Epub 2001/07/12. pmid:11446150.
  10. 10. NIH. Quality Assessment of Case-Control Studies. https://wwwnhlbinihgov/health-topics/study-quality-assessment-tools, accessed January 10, 2020.
  11. 11. Page MJ, Higgins JPT, Sterne JAC. Chapter 13: Assessing risk of bias due to missing results in a synthesis. In: Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions version 6.1 (updated September 2020). Cochrane, 2020. Available from
  12. 12. Webley WC, Salva PS, Andrzejewski C, Cirino F, West CA, Tilahun Y, et al. The bronchial lavage of pediatric patients with asthma contains infectious Chlamydia. Am J Respir Crit Care Med. 2005;171(10):1083–8. pmid:15735056.
  13. 13. Kopriva F, Szotkowska J, Zapalka M. Bronchial asthma and Chlamydia pneumoniae antibodies in children aged 4–8 years in Olomouc district. Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub. 2005;149(2):289–91. Epub 2006/04/08. pmid:16601774.
  14. 14. Teig N, Anders A, Schmidt C, Rieger C, Gatermann S. Chlamydophila pneumoniae and Mycoplasma pneumoniae in respiratory specimens of children with chronic lung diseases. Thorax. 2005;60:962–6. pmid:16143584
  15. 15. Dal Molin G, Longo B, Not T, Poli A, Campello C. A population based seroepidemiological survey of Chlamydia pneumoniae infections in schoolchildren. J Clin Pathol. 2005;58(6):617–20. Epub 2005/05/27. pmid:15917413; PubMed Central PMCID: PMC1770689.
  16. 16. Annagur A, Kendirli SG, Yilmaz M, Altintas DU, Inal A. Is there any relationship between asthma and asthma attack in children and atypical bacterial infections; Chlamydia pneumoniae, Mycoplasma pneumoniae and Helicobacter pylori. J Trop Pediatr. 2007;53(5):313–8. Epub 2007/05/31. pmid:17535826.
  17. 17. Wazir S, Kumar L, Sethi S, Sharma M. Seroprevalence of Chlamydia pneumoniae in asthmatic children from Northern India. Indian Pediatr. 2007;44(2):133–6. Epub 2007/03/14. pmid:17351305; PubMed Central PMCID: PMC17351305.
  18. 18. Nagy A, Keszei M, Kis Z, Budai I, Tolgyesi G, Ungvari I, et al. Chlamydophila pneumoniae infection status is dependent on the subtypes of asthma and allergy. Allergy Asthma Proc. 2007;28(1):58–63. Epub 2007/03/30. pmid:17390759.
  19. 19. Kazar J, Kovacova E, Gasparovic J, Cervenka J, Furkova K, Hornova J, et al. Antibody response to chlamydiae in children with asthma and respiratory illness. Folia Microbiol (Praha). 2011. Epub 2011/04/20. pmid:21503738.
  20. 20. Patel KK, Anderson EA, Salva PS, Webley WC. The prevalence and identity of Chlamydia-specific IgE in children with asthma and other chronic respiratory symptoms. Respir Res. 2012;13(1):32. Epub 2012/04/20. doi: 1465-9921-13-32 [pii] pmid:22512977.
  21. 21. Tutanç M, Gürkan FM, Yel S, Güneş A, Konca Ç, Bilen G. Seropositivity for Chlamydia pneumoniae and Mycoplasma pneumoniae in asthmatic children. Journal of Clinical and Analytical Medicine. 2015;6(3):353–7.
  22. 22. Park SJ, Lee YC, Rhee YK, Lee HB. Seroprevalence of Mycoplasma pneumoniae and Chlamydia pneumoniae in stable asthma and chronic obstructive pulmonary disease. J Korean Med Sci. 2005;20(2):225–8. Epub 2005/04/16. doi: 200504225 [pii]. pmid:15831991; PubMed Central PMCID: PMC2808596.
  23. 23. Rodríguez JA, Munoz FJ, Bellido JLM, Garcia Rodríguez JA. Prevalence of anti-Chlamydophila pneumoniae antibodies in patients with intrinsic asthma. Rev Esp Quimioterap. 2005;18:146–8. PubMed Central PMCID: PMC16130036.
  24. 24. Harju TH, Leinonen M, Nokso-Koivisto J, Korhonen T, Räty R, He Q, et al. Pathogenic bacteria and viruses in induced sputum or pharyngeal secretions of adults with stable asthma. Thorax. 2006;61(7):579–84. PubMed Central PMCID: PMC16517571. pmid:16517571
  25. 25. Paldanius M, Juvonen R, Leinonen M, Bloigu A, Silvennoinen-Kassinen S, Saikku P. Asthmatic persons are prone to the persistence of Chlamydia pneumoniae antibodies. Diagn Microbiol Infect Dis. 2007;59(2):117–22. Epub 2007/06/19. pmid:17572038.
  26. 26. Kocabas A, Avsar M, Hanta I, Koksal F, Kuleci S. Chlamydophila pneumoniae infection in adult asthmatics patients. Journal of Allergy. 2008;45:39–43. pmid:18259994
  27. 27. Torshizi AA, Mohammad T, Attaran D, Karamadin MK, Ghazvini K. Role of Chlamydia pneumoniae infection in asthma in northeast of Iran. Iran J Allergy Asthma Immunol. 2008;7:45–6. pmid:18322313.
  28. 28. Dejsomritrutai W, Siritantikorn S, Nana A. Asthma, bronchial hyper-responsiveness and Chlamydophila (Chlamydia) pneumonia infection in adult Thai population. J Med Assoc Thai. 2009;92 Suppl 2:S30–7. Epub 2009/07/01. pmid:19562983.
  29. 29. Specjalski K, Jassem E. Chlamydophila pneumoniae, Mycoplasma pneumoniae infections, and asthma control. Allergy Asthma Proc. 2011;32:9–17. PubMed Central PMCID: PMC21439159. pmid:21439159
  30. 30. Hahn DL, Schure A, Patel K, Childs T, Drizik E, Webley W. Chlamydia pneumoniae-specific IgE is prevalent in asthma and is associated with disease severity. PLoS ONE. 2012;7(4):e35945. pmid:22545149
  31. 31. Smith-Norowitz TA, Loeffler J, Huang Y, Klein E, Norowitz YM, Hammerschlag MR, et al. Chlamydia pneumoniae immunoglobulin E antibody levels in patients with asthma compared with non-asthma. Heliyon. 2020;6(2):e03512. Epub 2020/03/07. pmid:32140608; PubMed Central PMCID: PMC7052057.
  32. 32. Pasternak R, Huhtala H, Karjalainen J. Chlamydophila (Chlamydia) pneumoniae serology and asthma in adults: a longitudinal analysis. J All Clin Immunol. 2005;116(5):1123–8. pmid:16275386
  33. 33. Schmidt SM, Muller CE, Wiersbitzky SK. Inverse association between Chlamydia pneumoniae respiratory tract infection and initiation of asthma or allergic rhinitis in children. Pediatr Allergy Immunol. 2005;16(2):137–44. Epub 2005/03/25. doi: PAI229 [pii] pmid:15787871.
  34. 34. Zaitsu M. The development of asthma in wheezing infants with Chlamydia pneumoniae infection. J Asthma. 2007;44(7):565–8. Epub 2007/09/22. doi: 782170331 [pii] pmid:17885860
  35. 35. Zaitsu M. Does Chlamydia pneumoniae infection trigger to development of asthma in wheezy infants? J Asthma. 2009;46(9):967–8. Epub 2009/11/13. [pii]. pmid:19905928.
  36. 36. Foschino Barbaro MP, Resta O, Aliani M, Guido P, Izzo C, Logroscino C, et al. Seroprevalence of chronic Chlamydia pneumoniae infection in patients affected by chronic stable asthma. Clin Microbiol Infect. 2002;8(6):358–62. Epub 2002/06/27. pmid:12084104.
  37. 37. Gencay M, Rudiger JJ, Tamm M, Soler M, Perruchoud AP, Roth M. Increased frequency of Chlamydia pneumoniae antibodies in patients with asthma. Am J Respir Crit Care Med. 2001;163(5):1097–100. Epub 2001/04/24. pmid:11316642.
  38. 38. Hahn DL, Anttila T, Saikku P. Association of Chlamydia pneumoniae IgA antibodies with recently symptomatic asthma. Epidemiol Infect. 1996;117:513–7. pmid:8972677
  39. 39. Hahn DL, Peeling RW, Dillon E, McDonald R, Saikku P. Serologic markers for Chlamydia pneumoniae in asthma. Ann Allergy Asthma Immunol. 2000;84:227–33. pmid:10719781
  40. 40. Larsen FO, Norn S, Mordhorst CH, Skov PS, Milman N, Clementsen P. Chlamydia pneumoniae and possible relationship to asthma. Serum immunoglobulins and histamine release in patients and controls. APMIS. 1998;106(10):928–34. PubMed Central PMCID: PMC9833693. pmid:9833693
  41. 41. Mills GD, Lindeman JA, Fawcett JP, Herbison GP, Sears M. Chlamydia pneumoniae serological status is not associated with asthma in children or young adults. Int J Epidemiol. 2000;29:280–4. pmid:10817126
  42. 42. Nagy A, Kozma GT, Keszei M, Treszi A, Falus A, Szalai C. The development of asthma in children infected with Chlamydia pneumoniae is dependent on the modifying effect of mannose-binding lectin. J Allergy Clin Immunol. 2003;112:729–34. pmid:14564351
  43. 43. Routes JM, Nelson HS, Noda JA, Simon FT. Lack of correlation between Chlamydia pneumoniae antibody titers and adult-onset asthma. J Allergy Clin Immunol. 2000;105:391–2. PubMed Central PMCID: PMC10669864. pmid:10669864
  44. 44. Sirmatel F, Ustunsoy H, Sirmatel O, Akdemir I, Dikensoy O. The relationship between Chlamydia pneumoniae seropositivity and peripheral vascular diseases, acute myocardial infarction and late-onset asthma. Infection. 2003;31(5):367–8. Epub 2003/10/14. pmid:14556067.
  45. 45. Tuuminen T, Edelstein I, Punin A, Kislova N, Stratchounski L. Use of quantitative and objective enzyme immunoassays to investigate the possible association between Chlamydia pneumoniae and Mycoplasma pneumoniae antibodies and asthma. Clin Microbiol Infect. 2004;10(4):345–8. Epub 2004/04/03. [pii]. pmid:15059128.
  46. 46. von Hertzen L, Vasankari T, Liippo K, Wahlström E, Puolakkainen M. Chlamydia pneumoniae and severity of asthma. Scand J Infect Dis. 2002;34:22–7. pmid:11874160
  47. 47. Johnston SL, Blasi F, Black PN, Martin RJ, Farrell DJ, Nieman RB. The effect of telithromycin in acute exacerbations of asthma. N Engl J Med. 2006;354:1589–600. pmid:16611950
  48. 48. Johnston SL, Szigeti M, Cross M, Brightling C, Chaudhuri R, Harrison T, et al. Azithromycin for Acute Exacerbations of Asthma: The AZALEA Randomized Clinical Trial. JAMA Intern Med. 2016;176(11):1630–7. pmid:27653939.
  49. 49. Hiles SA, McDonald VM, Guilhermino M, Brusselle GG, Gibson PG. Does maintenance azithromycin reduce asthma exacerbations? An individual participant data meta-analysis. Eur Respir J. 2019: Epub 2019/09/14. pmid:31515407.
  50. 50. Steel HC, Theron AJ, Cockeran R, Anderson R, Feldman C. Pathogen- and host-directed anti-inflammatory activities of macrolide antibiotics. Mediators Inflamm. 2012;2012:584262. Epub 2012/07/11. pmid:22778497; PubMed Central PMCID: PMC3388425.
  51. 51. Grayston JT, Campbell LA, Kuo C-C, Mordhorst CH, Saikku P, Thom DH, et al. A new respiratory tract pathogen: Chlamydia pneumoniae strain TWAR. J Infect Dis. 1990;161:618–25. pmid:2181028
  52. 52. Burrows B, Martinez F, Halonen M, Barbee RA, Cline MG. Association of asthma with serum IgE levels and skin-test reactivity to allergens. New Engl J Med. 1989;320:271–7. pmid:2911321
  53. 53. Schmidt SM, Müller CE, Bruns R, Wiersbitzky SKW. Bronchial Chlamydia pneumoniae infection, markers of allergic inflammation and lung function in children. Pediatr Allergy Immunol. 2001;12:257–65. pmid:11737672
  54. 54. Wu L, Skinner SJM, Lambie N, Vuletic JC, Blasi F, Black PN. Immunohistochemical staining for Chlamydia pneumoniae is increased in lung tissue from subjects with chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2000;162:1148–51. pmid:10988144
  55. 55. Webley WC, Salva PS, Andrzejewski C, Cirino F, West CA, Tilahun Y, et al. The bronchial lavage of pediatric patients with asthma contains infectious Chlamydia. American Journal of Respiratory & Critical Care Medicine. 2005;171(10):1083–8. pmid:15735056. Language: English. Entry Date: 20050624. Revision Date: 20200624. Publication Type: Journal Article.
  56. 56. Boman J, Söderberg S, Forsberg J, Birgander LS, Allard A, Persson K, et al. High prevalence of Chlamydia pneumoniae DNA in peripheral blood mononuclear cells in patients with cardiovascular disease and in middle-aged blood donors. Journal of Infectious Diseases. 1998;178:274–7. pmid:9652454
  57. 57. Parrott GL, Kinjo T, Fujita J. A Compendium for Mycoplasma pneumoniae. Front Microbiol. 2016;7:513. pmid:27148202; PubMed Central PMCID: PMC4828434.
  58. 58. Yeh JJ, Wang YC, Hsu WH, Kao CH. Incident asthma and Mycoplasma pneumoniae: A nationwide cohort study. J Allergy Clin Immunol. 2016;137(4):1017–23 e6. Epub 2015/11/21. pmid:26586037.
  59. 59. Campbell LA, Hahn DL. Chlamydia pneumoniae Infections. In: Tan M, Hegemann JH, Sütterlin C, editors. Chlamydia Biology—From Genome to Disease. Norfolk, UK: Caister Academic Press; 2020. p. 31–58.
  60. 60. Hahn DL. Role of Chlamydia pneumoniae as an inducer of asthma. In: Friedman H, Yamamoto Y, Bendinelli M, editors. Chlamydia pneumoniae Infection and Disease. Infectious Agents and Pathogenesis. New York: Kluwer Academic/Plenum Publishers; 2004. p. 239–62.
  61. 61. Webley WC, Hahn DL. Infection-mediated asthma: etiology, mechanisms and treatment options, with focus on Chlamydia pneumoniae and macrolides. Respir Res. 2017;18(1):98. pmid:28526018; PubMed Central PMCID: PMC5437656.
  62. 62. Aaron SD, Vandemheen KL, FitzGerald JM, Ainslie M, Gupta S, Lemiere C, et al. Reevaluation of Diagnosis in Adults With Physician-Diagnosed Asthma. JAMA. 2017;317(3):269–79. pmid:28114551.
  63. 63. Dowell SF, Peeling RW, Boman J, Carlone GM, Fields BS, Guarner J, et al. Standardizing Chlamydia pneumoniae assays: Recommendations from the Centers for Disease Control and Prevention (USA) and the Laboratory Centre for Disease Control (Canada). Clin Infect Dis. 2001;33:492–503. pmid:11462186
  64. 64. Herland K, Akselsen J-P, Skjønsberg OH, Bjermer L. How representative are clinical study patients with asthma or COPD for a larger "real life" population of patients with obstructive lung disease? Respir Med. 2005;99:11–9. pmid:15672843
  65. 65. Travers J, Marsh S, Williams M, Weatherall M, Caldwell B, Shirtcliffe P, et al. External validity of randomised controlled trials in asthma: to whom do the results of the trials apply? Thorax. 2007;62:219–23. pmid:17105779
  66. 66. Hahn DL, Golubjatnikov R. Smoking is a potential confounder of the Chlamydia pneumoniae-coronary artery disease association. Arteriosclerosis & Thrombosis. 1992;12:945–7. pmid:1637792
  67. 67. Blasi F, Damato S, Cosentini R, Tarsia P, Raccanelli R, Centanni S, et al. Chlamydia pneumoniae and chronic bronchitis: association with severity and bacterial clearance following treatment. Thorax. 2002;57:672–6. pmid:12149525
  68. 68. Grayston JT, Aldous M, Easton A, Wang S-p, Kuo C-c, Campbell LA, et al. Evidence that Chlamydia pneumoniae causes pneumonia and bronchitis. J Infect Dis. 1993;168:1231–5. pmid:8228356
  69. 69. Hahn DL. A Cure for Asthma? What Your Doctor Isn’t Telling You, and Why. Durham, North Carolina: Peoples Pharmacy Press; 2013. 192 p.