Fraction of Exhaled Nitric Oxide Levels Are Elevated in People Living With Human Immunodeficiency Virus Compared to Uninfected Controls, Suggesting Increased Eosinophilic Airway Inflammation

Abstract

Background

Increased risk of asthma and chronic obstructive pulmonary disease has been reported in people living with human immunodeficiency virus (PLWH). Fraction of exhaled nitric oxide (FeNO) is a marker of eosinophilic airway inflammation. We assessed FeNO levels in PLWH and matched uninfected controls and investigated whether human immunodeficiency virus (HIV) status is independently associated with elevated FeNO.

Methods

FeNO was quantified by NIOX Vero and pulmonary function was assessed by spirometry in 432 PLWH from the Copenhagen Comorbidity in HIV Infection Study and in 1618 age- and sex-matched uninfected controls from the Copenhagen General Population Study. Elevated FeNO was defined as ≥25 parts per billion. Associations between FeNO and HIV status were adjusted for known potential confounders.

Results

Mean age of PLWH was 50.7 (standard deviation [SD], 11.1) years and 97.4% received combination antiretroviral therapy. PLWH had higher FeNO than uninfected controls (median, 17.0 [interquartile range {IQR}, 11.0–26.0] vs 13.0 [IQR, 9.0–19.0]; P < .001). Also, PLWH had a higher prevalence of elevated FeNO than uninfected controls (27.5% vs 12.3%; P < .001). This association remained after adjusting for age, sex, height, smoking status, use of airway medication, blood eosinophils, and immunoglobulin E (adjusted OR [aOR], 3.56 [95% CI, 2.51–5.04]; P < .001). Elevated FeNO was associated with self-reported asthma (aOR, 2.65 [95% CI, 1.66–4.24]; P < .001) but not with airflow limitation (aOR, 1.07 [95% CI, .71–1.62]; P = .745).

Conclusions

HIV status was independently associated with elevated FeNO, suggesting increased eosinophilic airway inflammation. The potential impact on chronic lung disease pathogenesis needs further investigation.

Introduction of combination antiretroviral therapy (cART) has increased survival of people living with human immunodeficiency virus (PLWH). However, even well-treated human immunodeficiency virus (HIV) infection is associated with chronic non-AIDS-related comorbidities including chronic pulmonary diseases [1], such as chronic obstructive pulmonary disease (COPD) [2] and spirometric airflow limitation [3–5]. While an association between airflow limitation and HIV has been well-described, less is known about the association between asthma and HIV infection. Only a few observational studies have been conducted in the cART era. These studies found a high prevalence of self-reported asthma and bronchodilator reversibility in PLWH compared to the background population [6–8].

The pathogenesis explaining these associations is not fully understood. Possible mechanisms include oxidative stress, immune cell activation, and chronic airway inflammation [9]. In this context, an important inflammatory process, increasingly recognized in the pathogenesis of obstructive pulmonary diseases, is eosinophilic airway inflammation [10, 11]. Fraction of exhaled nitric oxide (FeNO) level has been suggested as a biomarker to determine and quantify the degree of eosinophilic airway inflammation in patients with asthma [12, 13]. Thus, FeNO may provide important mechanistic information about the contribution of eosinophilic inflammation to pulmonary comorbidity in PLWH. Only a limited number of studies have investigated FeNO in PLWH, with conflicting results [14–16], and no prior studies have explored if well-treated adult PLWH have elevated FeNO levels compared to uninfected controls.

In the present study, we assessed FeNO levels in PLWH and matched uninfected controls from the general population. We hypothesized that HIV status is independently associated with elevated FeNO. Furthermore, we investigated whether elevated FeNO is associated with an increased risk of asthma and airflow limitation in PLWH and uninfected controls.

METHODS

Study Design and Population

In this cross-sectional study, we included PLWH enrolled in the Copenhagen Comorbidity in HIV Infection (COCOMO) Study, and age- and sex-matched uninfected controls from the Copenhagen General Population Study (CGPS). The COCOMO study is an ongoing prospective and noninterventional cohort study designed to investigate the prevalence, incidence, and pathogenesis of non-AIDS-related comorbidity in PLWH in the post-cART era [17]. The study was initiated in 2015 and includes >40% of PLWH in the greater Copenhagen area. Data collection for PLWH was performed from March 2015 until November 2016. The CGPS is an ongoing population-based prospective cohort study initiated in 2003 with >110 000 randomly selected participants from the greater Copenhagen area [18]. Uninfected controls were recruited from January 2013 through December 2016. All participants in COCOMO and CGPS completed questionnaires, underwent a physical examination, and provided blood for biochemical analyses. Questionnaires were reviewed in detail at the day of attendance by a healthcare professional together with the participant. Both studies have been approved by the Committee on Health Research Ethics of the Capital Region of Denmark (approval numbers H-8–2014–004 and H-KF-01–144/01) and the Danish Data Protection Agency. Written informed consent was obtained from all participants. The studies were conducted according to the Declaration of Helsinki.

Fraction of Exhaled Nitric Oxide

FeNO levels in the expiratory volume were measured by the portable hand-held device NIOX Vero (Aerocrine AB, Sölna, Sweden) and expressed in parts per billion (ppb) [19]. The apparatus has a lower detection limit of 5 ppb and a measurement range of 5–300 ppb. The measurements were done in a sitting position without the use of a nose-clip in accordance with the recommendations from the European Respiratory Society and the American Thoracic Society [20]. During the inspiration phase, individuals were required to inhale to their total lung capacity through the mouthpiece, which contains a protective filter, to avoid environmental containment. During the expiration phase, individuals were guided via an animated interface on the apparatus to maintain a correct and constant expiratory flow rate. If the individual failed to obtain or sustain a correct expiratory flow rate, FeNO level was not measured by the apparatus, and a repeated measurement was automatically required. All measurements were performed using the same procedure in COCOMO and CGPS. An elevated FeNO level was defined as ≥25 ppb, as recommended by the American Thoracic Society for interpreting FeNO levels in clinical practice [12]. While a FeNO level <25 ppb is used to indicate that eosinophilic airway inflammation is less likely, a FeNO level ≥50 ppb indicates that eosinophilic airway inflammation is likely [12, 13]. Thus, in main analyses, we used a cutoff value for FeNO of ≥25 ppb, and in sensitivity analyses a cutoff value of ≥50 ppb.

Spirometry

Spirometry was performed with an EasyOne Spirometer (ndd Medical, Zürich, Switzerland) with measurements of prebronchodilator forced expiratory volume in 1 second (FEV₁) and forced vital capacity (FVC). Airflow limitation was defined in 2 different ways as (i) FEV₁/FVC < lower limit of normal (LLN) or as (ii) FEV₁/FVC <0.70 according to the Global Initiative for Chronic Obstructive Lung Disease [21] with FEV₁ <80% of the predicted value (corresponding to COPD with moderate to very severe airflow limitation), as done previously [5]. Predicted values and LLN were calculated using lung function reference equations provided by the Global Lung Function Initiative [22].

Self-reported Outcomes

Information about asthma, allergy, use of airway medication, and tobacco smoking was obtained through self-report. Asthma was defined as an affirmative response to the question: “Do you have asthma?” Allergy was defined as an affirmative response to the question: “Does food, medicine, grass, flowers, or animal hair give you asthma, hay fever, or eczema?” Use of airway medication was defined as taking any kind of medication for asthma/bronchitis (including sprays/dry powders) daily or almost daily. Smoking status was defined as never, former, and current smoking.

Biochemistry

Plasma immunoglobulin E (IgE) concentrations were measured with the DiaS Immunoglobulin E FS assay (DiaSys, Holzheim, Germany). Blood eosinophil counts were measured with the ADVIA 120 Hematology System (Siemens Healthcare, Munich, Germany). In PLWH, the most recent CD4 T-cell count and HIV viral load were obtained from medical records.

Statistical Analysis

Uninfected controls were frequency matched with PLWH according to sex and 5-year age strata, aiming for 5 controls per PLWH. For men aged 20–55 years, it was not possible to identify 5 controls per HIV-infected individual due to differences in age and sex distribution between COCOMO and CGPS (Supplementary Figure 1). Differences in clinical characteristics were assessed using Student t test, Wilcoxon rank-sum test, and Pearson χ ² test. FeNO, blood eosinophils, and IgE concentrations were log-transformed to obtain a normal distribution. Multiple linear regression analysis was used to investigate the association between HIV status and FeNO levels. Using a stepwise approach, only variables associated with FeNO levels in unadjusted analysis were included in the multivariable adjusted model. The final model included HIV, age, sex, height, smoking status, use of airway medication, blood eosinophil count, and IgE. To determine risk factors associated with an elevated FeNO level (≥25 ppb), we performed logistic regression analysis and adjusted for the above-mentioned variables. Furthermore, we performed sensitivity analyses using a different cutoff for FeNO level (≥50 ppb) in logistic regression analysis. Finally, in explorative analyses, we investigated whether an elevated FeNO level was associated with asthma and/or airflow limitation in the entire study population and in analyses restricted to either PLWH or uninfected controls only. Using logistic regression analyses, we tested for statistical interaction between HIV infection and elevated FeNO level in their effects on the risk of asthma and airflow limitation, to determine whether HIV status modifies this association. Statistical analyses were performed using R software version 3.4.1.

RESULTS

A total of 432 PLWH and 1618 uninfected controls with FeNO measurements were included in the study (Table 1). PLWH were slightly younger and there was a higher prevalence of males and current smokers among PLWH than uninfected controls (Table 1). Most PLWH were receiving cART (97.4%), and the mean current CD4 T-cell count was  711 (standard deviation [SD], 275) cells/µL. Data on mean values are presented correct in this paragraph. Furthermore, PLWH had a lower FVC than uninfected controls (mean, 4.4 [SD, 1.0] L vs 4.7 [SD, 1.0] L; P < .001). No differences were observed in the prevalence of airflow limitation, self-reported asthma, allergy, or use of airway medication (Table 1).

FeNO Levels in PLWH and Uninfected Controls

PLWH had higher FeNO levels compared to uninfected controls (median, 17.0 [interquartile range {IQR}, 11.0–26.0] vs 13.0 [IQR, 9.0–19.0]; P < .001; Figure 1). For visualization of the entire distribution of FeNO stratified by HIV status, please see Supplementary Figure 2. A positive HIV status was independently associated with 39.1% (95% confidence interval [CI], 30.0–49.1; P < .001) higher FeNO level after adjusting for age, sex, height, smoking status, use of airway medication, blood eosinophil count, and IgE (Table 2). While age, male sex, height, blood eosinophils, and IgE were associated with higher FeNO levels, current smoking and former smoking were associated with lower FeNO levels (Table 2).

Association Between HIV Infection and Elevated FeNO

Elevated FeNO with ≥25 ppb was more prevalent in PLWH compared to uninfected controls (27.5% vs 12.3%; P < .001) (Figure 2). Likewise, in stratified analysis including only individuals with asthma, allergy, or current smoking, elevated FeNO was also more prevalent among PLWH compared to controls (Figure 2). We found HIV status to be associated with elevated FeNO in unadjusted model (odds ratio [OR], 2.71 [95% CI, 2.09–3.51]; P < .001; Table 3). This association remained statistically significant after adjustment for age, sex, height, smoking status, use of airway medication, blood eosinophils, and IgE (adjusted OR [aOR], 3.56 [95% CI, 2.51–5.04]; P < .001; Table 3). In sensitivity analyses, PLWH had a higher prevalence of FeNO level ≥50 ppb than uninfected controls (5.1% vs 1.6%; P < .001), and HIV status was independently associated with FeNO level ≥50 ppb in multivariable adjusted logistic regression analysis (aOR, 2.83 [95% CI, 1.25–6.43]; P = .013).

Association of FeNO, Asthma, and Airflow Limitation in PLWH and Uninfected Controls

Elevated FeNO was associated with asthma in the entire study population (aOR, 2.65 [95% CI, 1.66–4.24]; P < .001; Table 4). In contrast, elevated FeNO was not associated with risk of airflow limitation defined by either the LLN or the fixed criterion (Table 4). Similar results were found when limiting analyses to uninfected controls. A similar pattern was also found in analyses in PLWH only, but the association between elevated FeNO and risk of asthma did not reach statistical significance (Table 4). We found no evidence of interaction between HIV infection and FeNO level in the association with asthma or airflow limitation.

Association of HIV-related Disease Markers and FeNO

To determine whether HIV-related disease markers were associated with elevated FeNO, we conducted analyses in PLWH only (n = 432). We found no association of current CD4 T-cell count or a nadir CD4 T-cell count <200 cells/µL with elevated FeNO (aOR, .94 [95% CI, .83–1.06] per 100 cells/µL increase; P = .309) and (aOR, 1.29 [95% CI .71–2.36]; P = .406), respectively. Furthermore, neither a previous AIDS-defining condition nor time living with HIV was associated with elevated FeNO (Supplementary Table 1).

DISCUSSION

We measured FeNO levels in a large cohort of well-treated PLWH and matched uninfected controls from the same geographical area and found FeNO levels to be significantly higher in PLWH compared to uninfected controls. Also, we found HIV infection to be independently associated with elevated FeNO of ≥25 ppb and ≥50 ppb after adjusting for known potential confounders, suggesting increased eosinophilic airway inflammation. Elevated FeNO was associated with asthma but not with airflow limitation in the entire study population. A similar pattern was found in PLWH, albeit this did not reach statistical significance.

FeNO levels in treated PLWH have been assessed in 2 prior studies. The largest and most recent study was conducted in a cohort of HIV-positive children on cART in sub-Saharan Africa, and found decreased FeNO levels in 222 HIV-infected children compared to 97 uninfected controls [16]. However, approximately 25% of the cohort had a history of tuberculosis (TB) and uninfected controls were not matched on age and height, factors which are known to influence FeNO levels, especially in children [23]. Another small study found a decreased level of FeNO in 36 PLWH coinfected with TB compared to 46 healthy blood donors, but no difference was observed when compared to TB patients without HIV infection [15]. As TB has been found to reduce the activity of nitric oxide synthase and thus decrease the production of nitric oxide (NO) in the airways as an immune evasion mechanism by Mycobacterium tuberculosis [24], coinfection with TB may explain the observed decreased levels of FeNO found in this study. There are no previous studies that have compared FeNO in adult PLWH in the current cART era without TB to uninfected controls, making comparison to results from the present study difficult.

The mechanisms underlying higher FeNO levels in PLWH are unknown. Elevated FeNO has been observed in patients with asthma, allergy, and upper respiratory tract infections [11, 19, 25, 26]. On the other hand, smoking, which is usually more prevalent among PLWH, has consistently been associated with lower FeNO levels [12]. In the airways and lungs, NO is produced by the enzyme nitric oxide synthase (NOS) in a variety of cells including macrophages and endothelial cells [27]. The production of NO can be induced by proinflammatory cytokines such as interferon-γ and interleukin-1β released from macrophages and CD4 T-helper cells [28]. The high level of NO released by NOS may act as an immune effector molecule inhibiting viral replication, killing tumor cells, and eliminating various pathogens [29]. In vivo studies have shown that despite treatment with cART, alveolar macrophages harbor HIV virus and represent a potential viral reservoir [28]. In theory, this could lead to increased release of proinflammatory cytokines and thus explain higher levels of NO in the airways and lungs of PLWH compared to uninfected controls. This hypothesis is strengthened by experimental data demonstrating increased expression of inducible NO synthase accompanied by a significant production of NO in cultured human monocytes infected with HIV type 1 [30]. Also, it has been shown that serum nitrate level (the stable metabolite of NO) correlates with the amount of HIV DNA in peripheral blood mononuclear cells [31], again suggesting that HIV induces the production of NO by means of activated mononuclear phagocytes. Based on these findings, we speculate that elevated FeNO in PLWH is a consequence of the presence of HIV in the lungs.

Currently, the clinical use of FeNO is recommended for the diagnosis of eosinophilic airway inflammation, and hence the likelihood of inhaled steroid responsiveness in patients with asthma [10]. Furthermore, FeNO may also be a predictor of poor symptom control and severe exacerbations in asthma patients [26, 32]. Thus, elevated FeNO in PLWH may be more than just an epiphenomenon and may contribute to pulmonary comorbidity. Thus, we determined whether elevated FeNO was associated with measures of pulmonary function and self-reported asthma. We found elevated FeNO to be significantly associated with asthma in analyses conducted in the entire population. These findings were reproduced in analyses limited to uninfected controls, but not in analyses limited to PLWH. However, the odds ratio for asthma was only slightly lower in PLWH than in uninfected controls, and the lack of a significant association between elevated FeNO and asthma in PLWH may be due to lack of power in the stratified analyses, since only 7% of PLWH had self-reported asthma. Furthermore, PLWH are known to have elevated systemic inflammation that may impact FeNO [33]. Despite elevated levels of FeNO in PLWH, we found no difference in self-reported asthma.

We previously studied pulmonary comorbidity in PLWH and found HIV infection to be independently associated with a higher burden of self-reported respiratory morbidity [5], lower pulmonary function including a lower FEV₁ and FVC and airflow limitation defined by the fixed criterion [5], and higher lung clearance index [34], but not with the presence of emphysema [35]. In line with other studies, we found no association of elevated FeNO with airflow limitation defined by either the LLN or the fixed criterion [36]. Likewise, a previous study reported no association between airway obstruction and FeNO in HIV-infected children [16]. These results implies that FeNO probably has no direct pathophysiological role in airway obstruction, but rather is a marker of eosinophilic airway inflammation, as seen in persons with asthma [11].

Finally, in the present study, we found age, sex, height, blood eosinophils, and IgE to be significantly associated with elevated FeNO levels, whereas current smoking was associated with decreased FeNO levels. The latter may be due to high levels of NO in cigarette smoke that downregulate NOS [37]. These findings are consistent with several previous studies conducted in uninfected controls [12, 23, 37].

Our study has limitations. We were unable to make the predescribed 1-to-5 match in certain age strata and, instead, we adjusted for age and sex in the multivariable analyses. We did not have information on postbronchodilator spirometry, and so could not determine whether airflow limitation was reversible or irreversible. Some covariates including asthma and airway medication use were self-reported and recall bias cannot be excluded, although this would, presumably, have affected both PLWH and uninfected controls to the same extent. Also, we did not have information on inhaled corticosteroid use for each participant. We collected information about “use of airway medication,” which was included as a covariate in the multivariable adjusted analyses. However, we cannot exclude differences in use of inhaled corticosteroid between PLWH and controls. Finally, the COCOMO cohort consists predominantly of white males with well-controlled viral replication and our findings may not be generalizable to PLWH in other parts of the world. The strengths of the study include the large sample size, use of uninfected controls, and identical methods used across cohorts.

In conclusion, we found HIV infection to be independently associated with elevated FeNO. This finding implies that PLWH have more eosinophilic airway inflammation than uninfected controls. FeNO may contribute to the excess pulmonary comorbidity previously observed in PLWH, and we found elevated FeNO to be associated with self-reported asthma in the entire study population, but not in analyses restricted to PLWH. Prospective studies are warranted to better understand the long-term clinical impact of elevated FeNO in PLWH and to explore potential interventions.

Notes

Author contributions. R. F. T. was responsible for the study concept and statistical analysis and drafted the manuscript. N. L. P. H. was responsible for the study concept and statistical analysis and edited the manuscript. S. A. and Y. Ç. were responsible for concept, data collection, and edited the manuscript. M. G., A. D. K., D. M. K.-K., A. H. B., J. G., B. G. N., J. V., and J. L. were responsible for the concept and edited the manuscript. A. R. was responsible for concept, data collection, and statistical analysis and had content review and editing input. S. D. N. was responsible for the concept and data collection and edited the manuscript.

Acknowledgments. The authors thank all the study subjects for their participation. The authors thank the staff at the Department of Infectious Diseases at Rigshospitalet and at Hvidovre Hospital for their dedicated participation.

Financial support. This work was supported by Rigshospitalet Research Council, Region Hovedstaden; the Lundbeck Foundation; and the Novo Nordisk Foundation. The study was designed, conducted, analyzed, and written by the authors without involvement of any commercial party.

Potential conflicts of interest. R. F. T. has received grants from the Rigshospitalet Research Council and the Lundbeck Foundation and travel grants from Gilead. N. L. P. H. is supported by the Rigshospitalet Research Council. Y. Ç. has received personal fees from Sanofi Genzyme, Boehringer Ingelheim, and AstraZeneca, outside the submitted work, and is supported by the Lundbeck Foundation. M. G. has received grants from the Rigshospitalet Research Foundation and travel grants from Gilead. A. D. K. has received a grant from the Danish Heart Foundation and travel grants from Gilead. D. M. K.-K. has received travel grants from Gilead. A. H. B. is supported by the Lundbeck Foundation. J. V. has received personal fees for consulting and/or presenting from AstraZeneca, Boehringer-Ingelheim, Chiesi, GlaxoSmithKline (GSK), and Novartis, outside the submitted work, and is supported by the National Institute for Health Research Manchester Biomedical Research Centre. A. R. has received grants from the Aase and Ejnar Danielsen Foundation. S. D. N. has received unrestricted research grants from the Novo Nordisk Foundation, the Lundbeck Foundation, and the Rigshospitalet Research Council; has received travel grants from Gilead and GSK/ViiV; and has served on the advisory boards for Gilead and GSK/ViiV. All other authors report no potential conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes