Genetic profile and patient-reported outcomes in chronic obstructive pulmonary disease: A systematic review

Background Chronic Obstructive Pulmonary Disease (COPD) impacts differently on patients at similar grades, suggesting that factors other than lung function may influence patients’ experience of the disease. Recent studies have found associations between genetic variations and patient-reported outcomes (PROs). Identifying these associations might be fundamental to predict the disease progression and develop tailored interventions. This systematic review aimed to identify the genetic variations associated with PROs in COPD. Methods and findings Databases were searched until July 2017 (PROSPERO: CRD42016041639) and additional searches were conducted scanning the reference list of the articles. Two independent reviewers assessed the quality of studies using the Q-Genie checklist. This instrument is composed of 11 questions, each subdivided in 7 options from 1 poor-7 excellent. Thirteen studies reporting 5 PROs in association with genes were reviewed. Studies were rated between “good quality” (n = 8) and “moderate” (n = 5). The most reported PRO was frequency of exacerbations (n = 7/13), which was mainly associated with MBL2 gene variants. Other PRO’s were health-related quality of life (HRQOL) (n = 4/13), depressive symptoms (n = 1/13), exacerbation severity (n = 1/13) and breathlessness, cough and sputum (n = 1/13), which were commonly associated with other genetic variants. Conclusions Although a limited number of PRO’s have been related to genetic variations, findings suggest that there is a significant association between specific gene variants and the number/severity of exacerbations, depressive symptoms and HRQOL. Further research is needed to confirm these findings and assess the genetic influence on other dimensions of patients’ lives, since it may enhance our understanding and management of COPD.


Introduction
Chronic obstructive pulmonary disease (COPD) is a multifactorial, heterogeneous and progressive condition that affects 210 million people worldwide [1].Severity of COPD is usually classified according to the degree of airway obstruction (assessed with spirometry), nevertheless it has been acknowledged that people at similar grades of COPD report different disease impacts [1].These different reports among patients suggest that factors beyond lung function influence patients' experience of the disease.Indeed, upstream factors, such as the presence of specific genetic variants, have already been reported to play a role in this matter [2].For example, polymorphisms in SERPINA1 usually lead to a deficiency of the α1 antitrypsin, affecting 1-2% of all COPD cases [3].Additionally, the role of other candidate genes in the pathogenesis, comorbidities and outcomes of the disease have been studied [4].
Patient-reported outcomes (PROs) are a set of health outcomes directly reported by patients and may include symptoms (dyspnea, cough, pain, fatigue), exacerbation frequency and health status, among others [5].These outcomes are accepted as the most faithful representation of patients' perspectives of the impact of the disease and treatment benefits [6].The needed of assessing PROs has been considerably highlighted in the Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2017 update, which suggests that COPD classification should now be based on exacerbation frequency and patient's perception of their symptoms rather than on lung function only [1].
In the last years, it has become more evident that there is a strong association between PROs and genetics, namely in lung cancer [7], which may suggest that strong associations may also exist between genetics and PROs in COPD.However, a review of the known associations has never been conducted.The combination of genetics and PROs would be valuable to identify patients susceptible to PROs deficits, understand the diagnosis, predict disease progression and develop tailored and timely interventions [8].
Therefore, the focus of this systematic review was to synthetize the genetic variations associated with different PROs in COPD.

Search strategy
The systematic review protocol was registered at Prospective Register of Systematic Reviews (PROSPERO) (ref CRD42016041639).A comprehensive systematic search was conducted in May 2016 and weekly updates were performed until July 2017 in the following medical databases: PubMed , Scopus (1960Scopus ( -2016) ) and Web of Science .The PICOS (Populations, Intervention, Comparison, Outcome and Study Design) framework was used to develop literature search strategies, however Intervention (I) and Comparison (C) terms were omitted as they were not applicable to the present review [9].Accordingly, the search terms were based on a combination of the following keywords: [(COPD OR "chronic obstructive pulmonary disease" OR emphysema OR "chronic bronchitis") AND ("genetic associations" OR "genetic profile" OR "genetic analysis" OR gene) AND (dyspnea OR dyspnoea OR breathlessness OR fatigue OR cough OR depression OR anxiety OR "daily living" OR "quality of life" OR mood OR "well-being" OR "frequency of exacerbation" OR exacerbations OR "hospital admissions" OR "hospital length of stay" OR "acute exacerbations" OR "physical activity" OR "physical fitness" OR "physical function" OR "sputum production" OR phlegm OR pain OR "patient-reported outcomes" OR "patient-centered outcomes" OR "patient-centered outcomes")].The full search strategy can be found in supplementary material (S1 Table ).The reference lists of the selected articles were also scanned for other potential eligible studies.

Eligibility criteria
Studies were considered eligible if included adult patients (>18 years old) diagnosed with COPD and associated a genetic profile to one or more PRO.For the purpose of this systematic review, PRO were defined, according to the Cochrane Collaboration definition, as "reports coming directly from patients about how they feel or function in relation to a health condition and its therapy without interpretation by healthcare professionals or anyone else" [10].Studies were excluded if they were conducted in animals, were written in languages other than English, Spanish, French or Portuguese and did not differentiate chronic obstructive diseases (i.e., presented pooled data from several chronic obstructive diseases such as asthma, COPD and bronchiectasis).Book chapters, review papers, abstracts of communications on meetings, letters to the editor, commentaries to articles, unpublished work and study protocols were not considered suitable and, therefore, were also excluded from this study.This systematic review was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) (S2 Table) [9,11].

Quality assessment
Quality, internal validity and risk of bias of the included studies were assessed using the Quality of genetic association studies checklist (Q-Genie) [12].This instrument is composed of 11 questions to assess "rationale for study", "selection and definition of outcome of interest", "selection and comparability of comparison groups", "technical classification of the exposure", "non-technical classification of the exposure", "other source of bias", "sample size and power", "a priori planning of analysis", "statistical methods and control for confounding", "testing of assumptions and inferences for genetic analysis" and "appropriateness of inferences drawn from results".The Q-Genie checklist has 7 possible answers for each question (i.e., "1 (poor)", "2", "3 (good)", "4", "5 (very good)", "6", "7 (excellent)").The overall quality of studies is classified as "poor quality" if score is 35, "moderate quality" if score is >35 and 45, and "good quality" (>45), for studies having control groups.For studies without control groups, values for the parameters listed above are 32, >32 and 40, and >40, respectively [12].Two reviewers assessed the quality of studies independently.Disagreements were solved consulting a third reviewer.

Studies selection and data extraction
First, duplicates were removed and one reviewer performed the initial screening of title, abstract and keywords of studies based on the type of publication and relevance for the scope of the review.Then the full-text of each potentially relevant study was screened for content to decide its inclusion in the review.For each accepted study, one reviewer extracted the following data to a previously structured table: last author's name and year of publication, study design, sample characteristics (i.e., sample size, age, gender and COPD severity), PRO evaluated and outcome measures used, gene associated with the identified PRO and type of association between the PRO and the identified gene.Two independent reviewers further checked the extracted data for accuracy and completeness.Reviewers resolved disagreements by consensus.

Data analysis
Consistency of the studies quality assessment (performed by the two reviewers), was explored with the Cohen's kappa.The value of Cohen's kappa vary from 0 to 1 and can be interpreted as: i) 0.00-0.20:slight agreement; ii) 0.21-0.40:fair agreement; iii) 0.41-0.60:moderate agreement; iv) 0.61-0.80:substantial agreement; v) 0.81-1.00:almost perfect agreement [13].Statistical analysis was performed using IBM SPSS Statistics version 24.0 (IBM Corporation, Armonk, NY, USA).

Search strategy
Database search identified 1889 studies of potential interest.After duplicates removal, 1259 articles were analyzed for relevant content.From these, 1206 were excluded due to the following reasons: non-original articles (n = 575), absence of PRO/genetic associations (n = 387), non-specific for COPD (n = 122), studies conducted in animals (n = 98), studies written in other languages rather than English, Spanish, French or Portuguese (n = 24).The full-text of the remaining 53 potentially relevant articles was assessed and 40 articles were excluded.Reasons for exclusion included: absence of PRO (n = 19), absence of genetic association (n = 19), not specific for COPD (n = 1) and unavailability of the article even after contacting the authors (n = 1).In total, 13 articles were included, all published in English.A detailed diagram of the review process is presented in Fig 1.

Quality assessment
Articles scored between 39 and 55 in the Q-Genie checklist [12] (Table 1).Six articles were classified as studies without control group [14][15][16][17][18][19], from which five presented "good quality" and one was of "moderate quality".The remaining articles, were classified as "studies with control group", from which four were "moderate quality" and the remaining three of "good quality".Items with the lowest classification were the "selection and comparability of comparison groups" and "sample size and power".The agreement between the two independent reviewers was almost perfect (k = 0.83; 95% CI 0.29-1; p = 0.002).
Genetic variants and health-related quality of life.The impact of HO-1 and EPHX1 polymorphisms on treatments with N-acetylcysteine (NAC) was assessed [18,26].Better health-related quality of life, assessed with the SGRQ, was found in patients without the L allele (L-) of HO-1 gene, which is a (Gt) n polymorphism, than in those with the L allele (L+) relative to the activity score of SGRQ (SGRQ activity score: Baseline: 46.2±14,5; 16 weeks: 46.3±11.0;32 weeks: 46.7±12.2;48 weeks: 47.4±15.5;p = 0.02) [18].Additionally, patients having the slow activity group of the EPHX1 genotype (based on exon 3 polymorphism) also revealed better health-related quality of life than those having the fast activity group for the symptom score of      The impact of the Gly16Arg polymorphism of the ADRB2 gene on health-related quality of life HRQOL was also assessed.Significant differences were observed in all domains and total scores of SGRQ between both genetic groups (Arg/Arg and non-Arg/Arg), however only the impact and total scores were significantly different in patients with the Arg/Arg genotype (total score: -16.9 vs. -8.1,p = 0.005; impact score: -19.8 vs. -2.2p<0.001) [17].No significant associations were found when investigating impact of the ADRB2 polymorphism on treatment effect with budesonide/formoterol I(p = 0.909) and II(p = 0.648) on SGRQ [16].
Genetic variants and breathlessness, cough and sputum.No significant association was found between the Gly16Arg possible genotypes of ADRB2 and the scores for the BCSS scale (p>0.05)[16].

Discussion
This was the first systematic review to explore associations between genetics and PROs in patients with COPD.The 13 studies included reported on 12 genetic variations positively associated with 5 distinct PROs, i.e., exacerbation frequency and severity, depression, health related quality of life and symptoms (breathlessness, cough and sputum).
Most studies (n = 7/13) assessed the association of specific genetic variants with exacerbation frequency.This is an important remark since the frequency of exacerbations is strongly associated with patients' functional and physiological deterioration [27], reduced health related quality of life [28] and substantial morbidity and mortality [29].MBL2 was the gene mostly associated with frequency of exacerbations (3/13) [14,20,22].Several polymorphisms of MBL2 gene play important roles in the innate immunity as it encodes for mannose-binding lectine.The mannose-binding lectine is a pattern-recognition receptor that binds to the sugar structure presented in various microorganisms [30].Specific polymorphisms of the MBL2 gene have been found responsible for causing a decreased production of MBL (MBL-deficient genotype), and this has been associated with an increased risk of exacerbations.In fact, high MBL levels presented in serum have been associated with increased survival in COPD [14].Thus, MBL2 polymorphisms seem to be promising biomarkers to detect those with more susceptibility to exacerbations and good candidates for assessment with PRO-based approaches.
Polymorphisms in the GC and VDR genes causing deficits of vitamin D (associated with several comorbidities, such as osteoporosis or skeletal muscle dysfunction, in patients with COPD [31]), have also been connected to the frequency of exacerbations.Nevertheless, a careful interpretation of the literature is needed since both significant and non-significant associations between GC polymorphisms rs4588 and rs7041 or VDR (Bsm, Taql, Fokl) polymorphisms with frequency of exacerbations and lack of vitamin D have been reported [25,32,33].Additionally, many non-genetic factors may also lead to vitamin D deficiency, namely the absence of sun exposure, vitamin D retention on body fat or other social/cultural factors [31].Therefore, future studies are yet needed to enhance our understanding of the relationship between these polymorphisms, vitamin D deficiency and PRO in COPD.
Exacerbation severity has been significantly associated with a CCL1 allele for rs2282691 in one study [19].However, the authors' definition of exacerbation severity can be arguable, as they used death as endpoint.It is known that the severity of an exacerbation is not defined by mortality but rather by symptoms and number and length of hospitalizations in the most severe cases [34].Since the authors have recorded patients' main symptoms, it would be interesting to assess if the reported A allele for rs2282691 was also associated with those and if the T allele actually conferred protection to acute exacerbations of COPD, as suggested.Other option would be to explore the correlations between rs2282691 variant with specific instruments, such as the Exacerbations of Chronic Pulmonary Disease Tool (EXACT-PRO) [35] to assess the severity of exacerbations.This analysis would be essential to identify patients with higher predisposition for more severe exacerbations which would allow them to be targeted for more timely and directed monitoring and intervention.
Depression was found to be associated with the rs3794808 variant, affecting the SLC6A4 gene [21].It is known that depression presents a strong linkage to nicotine dependence [36] and SLC6A4 is strongly associated with the pathophysiology of tobacco use, namely at the level of serotonin reuptake.Therefore, it would be expected that specific genetic variants of this gene would play an important role on nicotine dependence and consequently depression in ex/current smokers [37].However, different SNPS of the SLC6A4 gene [38] and other genes such as THSD4, CHRNA, CYP2A6 [3] have also been associated with depression in COPD.Thus, it would be valuable to confirm those associations populations patients with COPD with different characteristics, such as smokers and non-smokers, to decide if future therapies should take these genes into consideration.Currently, the most effective therapy to combat anxiety and depression in patients with COPD is pulmonary rehabilitation (PR), which has been shown to be significantly effective by reducing the levels of depression and anxiety symptoms in patients with this disease [39].However, PR is an expensive therapy and access to it is highly limited [40].Therefore, genetics may be used to signal priority patients to PR, and thus optimize human and financial resources in managing COPD.
Four studies investigated ADRB2 [16,17], Ho-1 [18] and EPHX1 [26] association with health related quality of life.The Gly16Arg polymorphism of ADRB2 has been indicated as a risk factor for COPD [41].However, different results regarding its association with health-related quality of life emerged from our systematic review.Umeda et al., showed that patients with COPD and the Arg/Arg genotype presented better health related quality of life in treatments with tiotropium [17] whereas no significant associations was found in the study of Bleecker et al. for the same polymorphism and budesonide/formoterol treatment [16].The most obvious explanation is the substance used in the treatments, since other studies using other LABAs (long-acting b2-agonists) and LAMAs (long-acting muscarinic antagonists) have also presented no associations [42].As for Ho-1 and EPHX1 polymorphisms, both genes presented significant associations with SGRQ activity and symptoms sub-scores in patients with COPD that were treated with N-acetylcysteine (NAC) [18,26].Pharmacogenetics studies are of significant importance, since they investigate how genes affect a patient's response to drugs.This knowledge facilitates health-care by identifying patients that will respond differently to treatments.Future studies assessing health related quality of life may also include these genetic variants reported as being protective against COPD [43], since they may play an important role on patients' quality of life.
A final important aspect to emphasize is the ethnicity of the study populations.In this study we intended to summarize the genetic variants associated with PROS in patients with COPD.However, we observed that this systematic review included nine different ethnic groups from which the majority (n = 7/13) were conducted in Asian countries.This may explain the difficulty in obtaining similar results among different populations.For this reason, the ethnicity was also indicated in Table 2.
PROs are increasingly being understood as excellent instruments to translate a range of outcomes that spirometry cannot express, such as symptoms and patients' perspective of treatment [44].However, there is a massive number of PROMs (patient-reported outcomes measures) that were not found in this systematic review, and yet allow to assess other fundamental PROs such as mood, social and sexual life [6].Also, it was shown that genetics play a key role not only in the predisposition to the disease but also in common COPD-related comorbidities.Thus, further studies should be conducted to re-enforce the present knowledge and assess the genetic influence on other dimensions of patients' lives.

Limitations
This review has some limitations that need to be acknowledged.Firstly, the definition of PROs (although published) has not been used as a primary outcome in some of the included studies.Thus, few studies gave emphasis to it, giving priority to clinical outcomes which may have led to significant loss of studies and information.However, to minimize this problem we performed a meticulous choice of keywords to diminish the number of missed studies.Secondly, the constant changes in the definition of exacerbation in the GOLD [1] may also result in loss of studies over the years.We overcame this difficulty by enclosing studies which included participants that had reported exacerbations independently of the definition used at the time.Thirdly, studies used different methodologies to assess similar or different PROs and consequently, prevented the realization of meta-analysis.

Conclusions
This was the first systematic review to explore associations between genetics and PROs in COPD.Although a limited number of PROs have been successfully related to genetic variations, findings suggest that a significant association between specific genetic variants and the frequency and severity of exacerbations, health-related quality of life and depressive symptoms may exist.Thus, further research is needed to confirm these results and to assess the possibility of association of other genetic variants with other PROs in patients with COPD, since this may enhance our understanding and management of this disease.

Table 1 . Quality assessment scores for the selected studies based on the quality of genetic association studies (Q-Genie).
1-rationale for study; 2-selection and definition of outcome of interest; 3-selection and comparability of comparison group (if applicable); 4-technical classification of the exposure; 5-non-technical classification of the exposure; 6 other sources of bias; 7-sample size and power; 8-a priori planning of analysis; 9-statistical methods and control for confounding; 10-testing of assumptions and inferences for genetic analysis; 11-appropriateness of inferences drawn from results.All items have a maximum score of 7.Na-not applicable https://doi.org/10.1371/journal.pone.0198920.t001