Health-related quality of life is inversely correlated with C-reactive protein and age in Mycobacterium avium complex lung disease: a cross-sectional analysis of 235 patients

Background Mycobacterium avium complex (MAC) lung diseases generally cause chronic disease in immunocompetent hosts. Although a few studies have examined health-related quality of life (HRQL) in patients with MAC lung disease, there have been no large studies. This study aimed to evaluate HRQL and its correlation with clinical outcomes in MAC lung disease. Methods A cross-sectional study was conducted at Keio University Hospital to investigate the factors associated with HRQL in pulmonary nontuberculous mycobacterial diseases. MAC lung diseases were diagnosed according to the 2007 ATS/IDSA guidelines for nontuberculous mycobacterial diseases. The 36-item short form health survey (SF-36) was administered to assess clinical outcomes. Clinical variables included treatment status, latest haematological data, and bacterial smear and culture results. Results The SF-36 scores for the 235 patients (median age, 69 years; 45 men and 190 women) with MAC lung disease, except for the bodily pain and mental health subscale scores, were significantly lower than the Japanese population norms. In the multivariable analyses, current treatment for MAC and a positive sputum smear or culture within the past year were significantly associated with lower SF-36 scores. C-reactive protein (CRP) and age showed stronger inverse correlations with SF-36 scores. Conclusions HRQL, especially the physical component, was impaired in patients with MAC lung diseases; this appears to be related with current treatment status, positive sputum smear or culture within the previous year, and particularly CRP and age. Further studies including qualitative assessments are needed to investigate the efficacy of CRP as a marker for progression or treatment response in MAC lung disease. Trial registration Clinical trial registered with UMIN (UMIN000007964).


Background
The incidence of nontuberculous mycobacterial lung disease is increasing worldwide [1,2]. Mycobacterium avium complex (MAC) lung disease, the most common nontuberculous mycobacterial disease, generally causes chronic, slowly progressive disease in immunocompetent hosts [3].
Therapy involving multiple antimicrobials against MAC is generally effective because it decreases the bacterial load, resulting in a change from a positive to negative sputum culture. However, established MAC lung diseases are often incurable or recurrent, resembling other chronic diseases such as diabetes mellitus (DM), chronic obstructive pulmonary disease (COPD), or interstitial lung disease (ILD). Long-term treatment with multiple antimicrobial agents and their side effects frequently contribute to patient burden.
As an indicator of overall health status, patient-reported outcome measures representing health-related quality of life (HRQL), or the individual's satisfaction or happiness with aspects of life related to health, have become important in the current treatment guidelines for MAC lung disease, in addition to conventional measures of infection control or cure rate [3,4]. Furthermore, HRQL provides information for health management and policy decisions [5]. The 36-item Short-Form health survey (SF-36) version 2 is a general HRQL measure and has been used in different chronic respiratory diseases, including ILD [6], sarcoidosis [7], bronchiectasis [8], pulmonary tuberculosis [9], and COPD [10]. With MAC lung disease, only two studies reported impaired HRQL, as assessed using the SF-36 and St. George's Respiratory Questionnaire (SGRQ) [11,12], a respiratory disease-specific HRQL instrument originally designed for patients with COPD [13]. However, the correlations between HRQL and variables such as treatment status, sputum culture and smear results, and haematological data remain unknown.
This study aimed to evaluate HRQL, as assessed using the SF-36, and identify its clinical determinants in a large population of consecutive patients with MAC lung disease.

Study population
A cross-sectional study was conducted at Keio University Hospital with patients with pulmonary nontuberculous mycobacterial diseases (University Hospital Medical Information Network: UMIN000007964). The study protocol was approved by the Keio University Hospital ethics review board. Written informed consent was obtained from each patient.
We enrolled 285 patients aged ≥20 years who were diagnosed with nontuberculous mycobacterial diseases between May 2012 and February 2014 according to the 2007 American Thoracic Society/Infectious Disease Society of America guidelines for the diagnosis of nontuberculous mycobacterial diseases [3]. We excluded 31 patients who could not complete the HRQL questionnaire and 19 patients with nontuberculous mycobacterial diseases other than MAC. The final sample included 235 patients with MAC lung disease.
Health-related quality of life assessment All patients completed the SF-36 version 2 [14] and SGRQ in Japanese [13,15,16]. The SF-36 consists eight subscales: physical functioning, role-physical, bodily pain, general health, vitality, social functioning, role-emotional, and mental health. These eight subscales provide three summary scores, which were adjusted for Japanese patients: physical, mental, and role/social component summary scores [17]. All scores were transformed to fit a norm-based score for the general Japanese population, with a mean score of 50 and standard deviation of 10. Lower scores indicate poorer HRQL. The internal consistency of the SF-36 was assessed using Cronbach's α.
SGRQ scores (range, 0-100) were calculated for the total questionnaire as well as for each of the three components; the symptoms, activity, and impact components measure respiratory symptoms, impairment of physical activity, and social and psychological disturbances, respectively.

Assessment of clinical parameters
Demographic data included age at diagnosis, disease duration, body mass index (BMI), smoking status, underlying pulmonary diseases, comorbidities using the age-adjusted Charlson comorbidity index (CCI) [18], treatment status, and bacterial smear and culture results for M. avium and M. intracellulare at study enrolment. Haematological investigations and pulmonary function tests were performed after study enrolment. Treatment status was classified as never treated, previously treated, or currently treated. The lower respiratory tract specimens were cultured on egg-based solid media (Kyokuto Pharmaceutical Industrial Co., Ltd., Tokyo, Japan) or mycobacteria growth indicator tubes (Becton, Dickinson and Co., Sparks, MD, USA). The AccuProbe system (Gen-Probe Inc., San Diego, CA, USA) was used to identify MAC isolates. The COBAS AMPLICOR system (Roche Diagnostic Co., Ltd., Tokyo, Japan) and DNA-DNA hybridization test (Kyokuto Pharmaceutical Industrial Co., Ltd.) were used to classify MAC isolates as M. avium or M. intracellulare; unspecified MAC isolates were categorized as MAC. Sputum smears or cultures were defined as negative based on results obtained in the previous year. If patients did not cough up sputum, they were recorded as negative [19]. Pulmonary function tests were performed with patients with a stable condition using an electronic spirometer according to the American Thoracic Society guidelines [20].

Statistical analyses
Correlations between two continuous variables were analysed using Spearman's correlation coefficients. Comparisons were conducted using Student's t-tests between two groups and Tukey's tests among three groups. Analyses of covariance (ANCOVA), adjusted for age, sex, CCI, smoking status, BMI, and underlying pulmonary disease, were used to analyse group differences. To identify the relative contributions of factors to SF-36 scores lower than the Japanese population norms, factors significantly associated with lower norm-based SF-36 scores in the univariate analyses were entered into a stepwise forward and backward multiple regression model for multivariate analysis.
All P values were two-tailed; P < 0.05 was considered significant. Statistical analyses were conducted using JMP v11.0 (SAS Institute Japan Ltd, Tokyo, Japan).

Norm-based SF-36 scores
The norm-based SF-36 scores are shown in Fig. 1. Of the subscale components, the bodily pain, and mental health scores were within the range of the Japanese population norms; however, the physical functioning, role-physical, general health, vitality, social functioning, and role-emotional subscale scores were significantly lower than the Japanese population norms. The general health subscale represented the largest difference (difference, 6.4; 95 % CI, 5.1-7.7). For the summary scores, the physical and role/social component summary scores were significantly lower than the Japanese population norms.

Correlations between SF-36 subscale or component summary scores and clinical parameters, physiological parameters, and SGRQ scores
The internal consistency of the SF-36 was good (Cronbach's α, 0.90-0.92). The correlations between the SF-36 subscale or component summary scores and pulmonary function tests and SGRQ scores are shown in Table 2. All SF-36 subscale scores were significantly correlated Pulmonary function showed stronger correlations with the physical functioning, role-physical, and general health subscale scores and the physical component summary score than the other subscale scores or component summary scores. Age was significantly correlated with all SF-36 scores except the general health, vitality, and mental health subscale scores and the mental component summary score; the disease duration was poorly correlated with all SF-36 scores. BMI had a significant but weak correlation with the SF-36 scores, excluding the physical component summary score. Age-adjusted CCI had a significant and moderate correlation with the physical functioning and role-physical subscale scores and the physical component summary scores.

Correlations between SF-36 subscale or component summary scores and haematological data
Correlations between the SF-36 subscale or component summary scores and haematological data are shown in Table 3. Haemoglobin (Hb) and C-reactive protein (CRP) levels were significantly correlated with the physical functioning, role-physical, general health, vitality, social functioning, and role-emotional subscale scores and the physical and role/social component summary scores; CRP was strongly correlated with the physical functioning subscale score. Alkaline phosphatase, sialylated carbohydrate antigen KL-6, and white blood cell count were moderately correlated with physical component summary score. Hb was moderately correlated with the physical functioning and role-physical subscale scores and physical component summary scores.

Comparisons between SF-36 subscale or component summary scores using ANCOVA
The physical functioning, role-physical, general health, and social functioning subscale and physical component summary scores significantly differed between the neverand currently-treated groups and between the previously-and currently-treated groups, but not between the never-and previously-treated groups (Fig. 2a). The physical functioning, role-physical, general health, vitality, and role-emotional subscale scores and the physical and role/social component summary scores significantly differed by sputum smear or culture results (Fig. 2b, c).

Multivariate determinants for SF-36 subscale or component summary scores
The stepwise multiple regression analysis included age, sex, age-adjusted CCI, BMI, smoking status, disease duration, underlying pulmonary diseases, treatment status (currently treated vs previously/never treated), positive sputum smear or culture, pulmonary function, Hb levels, and CRP levels (  we also identified several factors, including CRP levels and age that were strongly associated with poorer HRQL. A number of previous studies have demonstrated impaired HRQL in various chronic diseases, including DM; patients with DM had lower scores for all eight SF-36 subscales compared with US population norms [21] and a higher risk for low physical component summary scores compared with older Japanese adults without DM [22]. In stroke patients, neurological impairments were associated with functional disability shortly after the stroke; lower physical and mental component summary scores were present at 3 months and 1 year after the stroke [23]. In addition, in the general population, common chronic conditions (e.g., ischemic heart disease, DM, and arthritis) notably impacted the physical functioning, general health, and bodily pain subscales, respectively, in a previous study. Furthermore, lower SF-36 physical component summary scores have been reported for patients with arthritis, chronic lung disease, or congestive heart failure [24]. The severity of chronic lung diseases, including chronic bronchitis, emphysema, and asthma, are reportedly significantly correlated with all eight SF-36 scales, reflecting respiratory symptoms [25,26]; the physical component summary score is primarily affected in COPD [27], asthma [28], ILD [6], and sarcoidosis [7], while bronchiectasis impacts both the physical and mental component summary scores [8]. To the best of our knowledge, only two studies have evaluated HRQL in patients with MAC lung disease. Impaired SGRQ-assessed HRQL was associated with high-resolution computed tomography findings [12]. Furthermore, in a study based in Canada, all of the subscale and component summary scores were impaired [11]; comparatively, the bodily pain and mental health subscale scores, and the mental component summary scores were within the population norm ranges in the present study. Pulmonary function was better, and there were fewer patients being currently treated in the present study, which might explain the differences in the results, in addition to differences in nationality or race.
In the present study, there were differences in HRQL, as evaluated using SF-36, between the treatment statuses. Patients currently receiving medical treatment had lower HRQL than the other treatment groups, while there were no differences in HRQL between the neverand previously-treated groups. These results suggest that multiple antimicrobial therapies against MAC might have a significant negative impact on HRQL, although patients who are currently receiving treatment tend to have more severe MAC lung disease. Sputum status, especially in culture conversion, has been used classically to evaluate infection control in MAC lung disease [29][30][31]. Notably, our data indicate the ability of sputum status to measure HRQL in patients with MAC lung disease.
To the best of our knowledge, our study is the first to demonstrate strong correlations between SF-36 scores and multiple clinical factors including patient characteristics, sputum culture status, pulmonary function, and haematological data. Age has already been reported as an important factor negatively influencing the SF-36 score [32]; given the aging population, this correlation is likely to become more important [33]. Furthermore, our study revealed that CRP might reflect HRQL-associated disease activity in patients with MAC lung disease. CRP is a biomarker that is elevated in acute and chronic inflammation such as that in infectious disease or noninfectious chronic diseases. Previous studies in chronic disease have demonstrated the ability of CRP to predict cardiovascular events in patients with arteriosclerotic diseases [34], treatment failure in tuberculosis [35] or human immunodeficiency virus infection [36], and disease progression in ILD with systemic scleroderma [37]. CRP might be a better and more convenient marker than a physiological parameter and could potentially be a marker of disease progression or treatment response.
Our study has several potential limitations. First, the cross-sectional design creates challenges in determining causal relationships, particularly regarding the influence of treatment. Second, by including only patients who could complete the SF-36, SGRQ, and pulmonary function tests, we might have excluded patients with more severe disease. However, we believe the present findings also apply to patients with severe disease because the pulmonary function and some of the haematological results were significantly correlated with HRQL.

Conclusions
In conclusion, patients with MAC lung disease have significantly lower HRQL than the Japanese population norms, indicating impaired physical performance and social activities. These impairments appear to be related