Pulmonary complications have been associated with human immunodeficiency virus (HIV) infection from the beginning of the acquired immunodeficiency syndrome (AIDS) epidemic.1–4 Chronic lung diseases and other comorbid non-AIDS conditions are now associated with a relatively increased number of deaths in persons with HIV as life expectancy increases because of improved HIV treatment and education.5–9 Numerous publications have provided pictorial essays illustrating common radiographic findings related to the lung and HIV infection or AIDS.2,10–20 Some of the most common radiographic findings in both the preantiretroviral therapy (ART) era and ART era include emphysema,21–27 focal air trapping,28 pulmonary infiltrates,1,10,23,29 alveolar consolidation,11,30,31 nodules,10,11,29,31–38 ground glass opacities,11,32,39,40 cavitary lesions,32,35 and bronchial abnormalities.11,24,34,41 Several early studies described precocious emphysema present in people with HIV infection or AIDS.21,22,25
The improved survival of HIV-infected individuals may have shifted the spectrum of chest computed tomographic (CT) findings to include more manifestations of chronic lung diseases, such as chronic obstructive pulmonary diseases (COPD)42–47 and interstitial inflammation or fibrosis, but only a few large-scale studies of chest CT findings have been performed in the current era.46,48–52 Furthermore, to our knowledge, only 2 studies performed quantitative CT analyses.27,52 A quantitative approach may be more sensitive to detect correlations between lung abnormalities on radiographic images and biological/physiological measures and to detect lung abnormalities at an earlier stage compared to subjective visual image review.
In our current multicenter study, we quantitatively analyzed chest CT images to determine the prevalence of findings indicative of emphysema and pulmonary fibrosis and to investigate the relation between demographics, lung function, biological markers of HIV infection, and CT image metrics associated with lung disease.
Subjects were participants in the National Heart, Lung, and Blood Institute (NHLBI)-sponsored Lung-HIV study from 3 Lung-HIV clinical centers: Ohio State University, University of Pittsburgh, and the University of Washington.53,54 The University of Pittsburgh center had 2 subcenters at the University of California, Los Angeles and the University of California, San Francisco. The University of Washington center recruited participants from 4 Veterans Affairs centers located in Atlanta, GA; Bronx, NY; Houston, TX; and Los Angeles, CA. The Institutional Review board at each clinical center, the Lung-HIV Data Coordinating Center (Clinical Trials and Surveys Corp, C-TASC), the Lung-HIV Data Safety and Monitoring Board, and the NHLBI approved the study protocol, and all participants signed written informed consent.
Each site recruited HIV-infected adults (≥18 years of age) who were without acute respiratory illness at the time of study enrollment. Subjects were not, however, required to have chronic pulmonary disease to participate. At the Ohio State University site, all subjects enrolled were current cigarette smokers, whereas subjects were eligible to enroll in the other 2 centers regardless of smoking status. The participants completed standardized questionnaires to ascertain demographics, smoking history, illicit drug use, and medical history.53,54 Participants were defined as never smokers if they reported smoking <100 lifetime cigarettes; as current smokers if they had smoked within the past 12 months; and as former smokers if they had quit more than 12 months before enrollment. Smoking pack-years were calculated based on average number of packs of cigarettes smoked per day and number of years smoked. Current CD4 cell counts (cells per microliter) and HIV viral load (copies per milliliter) were obtained within 12 months of pulmonary function testing.
Pulmonary Function Testing
Pulmonary function tests were performed according to the American Thoracic Society/European Respiratory Society (ATS/ERS) guidelines at each site's clinical or research pulmonary function laboratory and included prebronchodilator and postbronchodilator spirometry and single breath carbon monoxide diffusion capacity of the lung (DLco).55,56 Spirometry and DLco–predicted normal values (or percent predicted) were computed based on the Hankinson57 and Neas58 formulas, respectively. Spirometry measures of lung function evaluated included percent predicted forced expiratory volume in 1 second (FEV1%), percent predicted forced vital capacity (FVC%), and the ratio of FEV1 to FVC (FEV1/FVC). Pulmonary function testing data have been previously reported for a subset of this cohort.53,54
The CT examinations were performed at 8 clinical sites using 7 scanner models. The CT protocol was standardized across the different scanners to control for scanner differences (see Table S1, Supplemental Digital Content, http://links.lww.com/QAI/A768). Quality assurance measures were implemented to evaluate each scanner and each CT scan for protocol compliance (see Supplemental Digital Content, http://links.lww.com/QAI/A768). The CT data were acquired without radiopaque contrast at a moderate radiation exposure (100 mAs) with the participants in a supine position and holding their breath at end inspiration. The image thickness was 0.625, 0.750, and 0.900 mm for General Electric, Siemens, and Phillips scanners, respectively.
CT Image Analysis
The CT data were analyzed using a standardized approach at the University of Pittsburgh by an image analyst who was blinded to all clinical information. The lung was first segmented from the CT images using University of Pittsburgh software.59 Then, we used the density mask technique60 to quantify the percentage of the lung voxels associated with emphysema using a threshold of −950 Hounsfield units (HU) (Perc950), which is a widely accepted threshold to quantify emphysema on thin-section CT images.61,62 To quantify the presence of interstitial inflammation, fibrosis, or fibrosis-like changes in the lung, we calculated the percentage of the lung voxels with a computed attenuation between −600 and −250 HU or so-called high attenuation areas (PercHAA).63 This HU value range captures fibrosis-like changes that are more dense than healthy parenchyma (approximate average, −750 HU) and less dense than water (0 HU). Both metrics were computed from images reconstructed using either “standard” kernel, “B31f” kernel, or “B” kernel of the General Electric, Siemens, and Phillips scanners, respectively. We classified CT scans as depicting trace levels of “emphysema” and fibrosis-like changes based on 2.5% and 5.0% cutoffs, respectively. These levels are intended to capture early structural changes in lung anatomy that are associated with disease and are mostly likely not clinically significant (eg, change patient care). For completeness, we also included emphysema and fibrosis cutoffs of 5.0% and 10.0%, respectively.
Subject demographics were described using mean and SD, median and interquartile range (IQR), or numbers and proportions as appropriate. Pearson correlation coefficients (PCC) were computed to quantify the association between quantitative emphysema (Perc950), quantitative fibrosis-like changes (PercHAA), age, smoking history (pack-years), current CD4 cell count, HIV viral load, and lung function with all variables treated as continuous variables. Stepwise multivariable linear regression analyses were performed with either Perc950 or PercHAA as the dependent variable to further investigate their relation with current CD4 cell count and HIV viral load. Based on both biologic plausibility and previous literature examining lung disease in HIV, we examined a number of potential confounders in stepwise models to determine the effect of adding different covariates en bloc to our primary predictors of interest, the current CD4 cell count and HIV viral load. First, demographic variables (age, race, and smoking history) were included in model 1. Next, ART status was added (model 2), and then drug use (marijuana or other illicit drugs) (model 3). Finally, the history of pertinent lung disease (bacterial pneumonia and Pneumocystis pneumonia) was added to the analysis (model 4). In both correlation and multivariable analyses, Perc950 and PercHAA were log transformed, continuous variables. In the multivariable analysis, current CD4 cell count was dichotomized to less than 500 cells per microliter and greater than or equal to 500 cells per microliter, and HIV viral load was dichotomized to less than 500 copies per milliliter or greater than or equal to 500 copies per milliliter based on the distribution of the data. Smoking status was defined as current, former, or never. Race was defined as white, black, or other. In all statistical tests, a P value less than 0.05 was considered significant.
Lung HIV CT Study Cohort
The study cohort included 510 HIV-infected individuals; of whom, 350 (69%) were on ART at the time of enrollment (Table 1). The cohort was 81% male with a mean age (SD) of 48.9 (±9.5) years. The cohort was racially diverse with 255 (50%) non-Whites. Overall, 77 (15%) of participants were life-long never smokers, and 325 (64%) were current smokers with a median (IQR) smoking history of 18.4 (29.1) pack-years. Injection drug use, marijuana use, previous bacterial pneumonia, and previous Pneumocystis pneumonia were prevalent. The median (IQR) current CD4 cell count and median (IQR) HIV viral load were 466 (398) cells per microliter and 48 (220) copies per milliliter, respectively. The number of participants with a current CD4 cell count less than 200 cells per microliter and an HIV viral load greater than 500 copies per milliliter was 67 (13%) and 199 (39%), respectively. The mean (SD) values of postbronchodilator FEV1%, FVC%, and FEV1/FVC were 95.1 (16.5), 98.6 (14.4), and 77.0 (10.0), respectively, and the DLco% was 70.6 (17.5).
Emphysema Quantified on CT Exam
The prevalence of trace (Perc950 >2.5%) or greater levels of emphysema in the study cohort was 25.1% (Table 1) and were 9.2% for Perc950 >5.0. There was a positive correlation between emphysema and both age (PCC = 0.283) and smoking history (PCC = 0.164) (Table 2). Neither current CD4 cell count nor HIV viral load was significantly correlated with quantified emphysema. In multivariable analyses using the Perc950 to reflect degree of emphysema as a continuous outcome, there was not a significant relationship between quantitative emphysema and current CD4 cell count, HIV viral load, or ART use in the 4 linear regression models (see Tables S2-S5, Supplemental Digital Content, http://links.lww.com/QAI/A768).
There was a significant positive correlation between lung function and emphysema with FVC% and a negative correlation with FEV1/FVC and quantitative emphysema (Table 2). The association between FEV1/FVC and quantitative emphysema remained when controlling for age and smoking history (see Table S6, Supplemental Digital Content, http://links.lww.com/QAI/A768). The correlation between quantitative emphysema and DLco did not reach statistical significance.
Fibrosis Quantified on CT Exam
Overall, 29.4% of the study cohort had at least a trace level (PercHAA >5.0%) of fibrosis-like changes in the lung (Table 1) and 5.7% of the cohort had PercHAA >10.0%. Correlations between findings of fibrosis were statistically significant with increasing HIV viral load (PCC = 0.210) and decreasing age (PCC = −0.239) but were not significantly correlated with pack-years or current CD4 cell count (Table 2). Age, smoking status, and ART had a significant relation with fibrosis-like changes in the multivariable regression models 1–4 (see Tables S7 and S8, Supplemental Digital Content, http://links.lww.com/QAI/A768). In multivariable analyses, HIV viral load and smoking status had a significant positive correlation with fibrosis-like changes in the lung (model 1), whereas age had a significant negative relation with fibrosis-like changes (Table 3). When ART was added to the model, age had significant negative correlation with fibrosis-like changes, and participants on ART had significantly lower levels of fibrosis-like changes. These associations remained true as other variables were added to the models (see Tables S9 and S10, Supplemental Digital Content, http://links.lww.com/QAI/A768).
Lung function measures (FEV1% and FVC%) had a significant negative correlation with fibrosis-like changes in the lung (Table 2). The association with FVC% remained when controlling for smoking history (see Table S11, Supplemental Digital Content, http://links.lww.com/QAI/A768). The correlation between fibrosis-like changes and DLco did not reach statistical significance.
Only 1.2% (6/512) of the subjects were positive for both quantitative emphysema and fibrosis-like changes.
This study is one of the few to investigate the relation between quantitative assessments of CT image metrics associated with lung disease and demographics, biological markers of HIV infection, and pulmonary function. As expected from published studies, we found that a significant number (25%) of HIV-infected individuals have evidence of at least early emphysema. In contrast, an unexpected finding was that a similar proportion (29%) also had fibrosis-like changes in the lung on quantitative analysis as pulmonary fibrosis has been a relatively uncommon chronic lung disease in HIV-infected persons. In addition, higher HIV viral levels were linked to fibrosis-like changes, but not to emphysema.
A novel finding from our study is the significant correlation between HIV viral load and quantified fibrosis-like changes in the lung (PercHAA) depicted on CT images. The histologic correlate of these areas of high computed attenuation (HAA) on CT images has not been established, but it has been hypothesized that regions of increased computed attenuation may indicate areas of the lung affected by subclinical interstitial lung disease (ILD) or early pulmonary fibrosis.63 In the Multi-Ethnic Study of Atherosclerosis (MESA) study, Lederer et al63 performed CT scans for scoring coronary calcium in a population of 2563 adults without clinical cardiovascular disease or airflow obstruction and found that increased percentage of high attenuation areas, defined as regions with pixel values between −600 and −250 HU, increased for each 10 pack-year cigarette smoking history. When PercHAA was correlated with visual changes on CT scan, the most common findings were ground glass opacities or atelectasis, reticular abnormalities, and a possible usual interstitial pneumonitis pattern.
Of note, a recent population-based cohort study in an HIV-uninfected population has shown that CT evidence of subclinical ILD is an underrecognized phenomenon.64 Further understanding of the relation between HIV viral load and CT findings suggesting subclinical ILD may inform our understanding of early events in the etiology and pathogenesis of ILD in the general population. Because the precise etiology and natural history of fibrosis-like changes in the lung have not been fully elucidated, the significance of our findings and the relationship between these CT image changes and viral load in HIV remain speculative.
Although areas of increased computed attenuation may indicate early ILD, it is also possible that such changes may reflect inflammatory changes, which are early events in the pathogenesis of other HIV-related pulmonary disorders. For example, recent data in the general population have demonstrated a significant correlation between fibrosis-like changes and biomarkers of pulmonary vascular endothelial injury, including intercellular adhesion molecule 1.65 Pulmonary vascular abnormalities are well described in the HIV-infected population, and HIV may have direct effects on the vascular endothelium. Evidence exists that HIV Nef, an accessory HIV protein, may directly affect endothelial function, reducing endothelial vasorelaxation and endothelial nitric oxide synthase expression while increasing oxygen-free radical production.66 Alternatively, there may be direct effects of drugs on pulmonary inflammatory pathways. For example, antiretroviral agents may directly increase leukocyte adhesion to vascular endothelium in vitro.67 Whether HIV-infected individuals with increased findings of fibrosis are at increased risk of developing “vascular-based” clinical abnormalities, such as pulmonary diffusion impairment or pulmonary hypertension, remains speculative. Finally, it is possible that individuals with areas of increased computed attenuation on CT images may have increased colonization of lung pathogens.
As expected, there was a significant correlation between quantitative findings of emphysema (Perc950) and both age and smoking history (Table 2). A number of studies in both pre-ART and ART eras have demonstrated a link between HIV status and an increased risk of developing COPD/emphysema.21,25,27,28,32,46,48,49 Although evidence exists that this phenomenon may be related to an interaction between HIV infection and smoking, we did not find a significant correlation between CD4 cell count or HIV viral load and Perc950. Our findings are consistent with Samperiz et al52 who reported no relation between quantitative assessment of emphysema on CT scan and either HIV viral load or CD4 count. However, they reported that the Center for Disease Control HIV category C was a risk factor for quantitative emphysema and a modest emphysema prevalence (10.5%) using a low emphysema cutoff (greater than 1% of voxels with a value less than −950HU). Three recent studies of HIV-infected cohorts46,48,52 observed that greater than 30% of the subjects had some evidence of emphysema by visual CT image assessment, which is somewhat higher than our findings of 25.1% by quantitative analysis. However, all 3 studies reported some significant relation between either nadir or current CD4 count and emphysema.
In the recent studies of emphysema in the HIV-infected population mentioned above, there seems to be some discordance between visual and quantitative assessments of emphysema prevalence. This discordance has long been discussed in the literature with the reported agreement by kappa statistic ranging from 0.12 to 0.48.68–73 One hypothesis is that the poor-to-moderate interreader agreement for visual emphysema assessment contributes to the discordance between visual and quantitative assessments with reported agreement kappa ranging from 0.18 to 0.63.68–73 Some investigators postulated that human observers tend to overestimate emphysema compared with quantitative assessment in subjects with low-level emphysema, but that reader agreement improves as emphysema severity increases.68 However, poor interreader agreement was observed in both subjects without COPD and subjects with severe COPD.71 One possible component of the discordance is the selection of an emphysema threshold (eg, −950HU) and cutoff for classifying emphysema (eg, 5% of voxels below the threshold). Investigators have used emphysema cutoff values ranging from 1% to 10% often based on empirical data, lung function data, or clinical outcomes (eg, mortality).52,71,74−76 We presented data for cutoff values of 2.5% and 5.0% for emphysema and fibrosis, respectively, and believe those values appropriately capture emphysema and fibrosis-like changes in our cohort based on preliminary visual assessment (unpublished data) and recent reports of other HIV studies. Although the visual and quantitative assessments of emphysema should be concordant, some investigators postulate that the 2 metrics are complementary with different strengths and weaknesses.71 Exactly how the 2 metrics complement each other and how to integrate the 2 are not quite clear.
A potential limitation to our study is the application of a predefined threshold to perform the quantitative assessment of emphysema and fibrosis-like changes in the lung on CT data obtained from different CT scanners. This application may or may not have been appropriate as there is no universal threshold or adjustment to the threshold to account for difference in CT scanners and protocols, but the threshold approach is widely used to quantify emphysema on CT images. Additionally, the values used to dichotomize into categories of emphysema (2.5% or 5.0%) and fibrosis-like changes (5.0% or 10.0%) have not been universally validated. However, we believe that our effort to standardize the CT protocols permits an analysis that would detect significant first-order effects but may be insufficient to detect higher-order more subtle effects limited by the size of our cohort. Another limitation is the cutoff selected for CD4 cell count and HIV viral load in the multivariable models, which can be debated regarding the optimal value. However, our initial multivariable analyses treated CD4 cell count and HIV viral load as continuous variables, but the data did not reach statistical significance (data not shown). Additionally, perhaps, nadir CD4 cell count and zenith HIV viral load levels are more appropriate in the context of chronic lung disease but were not available for these analyses. Finally, our cohort had a high prevalence of smoking, particularly given that all patients from the Ohio State University site were current smokers. Nonetheless, our multisite study may be more representative of the general HIV-infected population, and we had substantial geographic and racial diversity in our population.
In summary, quantitative CT assessment of the lungs shows that emphysema and fibrosis-like changes were both common in a multicenter cohort of HIV-infected individuals. Fibrosis-like changes were associated with HIV viral load when controlling for age, race, smoking, and ART. In contrast, quantitative assessment of emphysema was not associated with HIV markers. Further investigation and extended follow-up of participants may shed light on our understanding of HIV-associated pulmonary complications and also provide new insight into events important in the development of early ILD.
The authors acknowledge and thank Dr. Hannah Peavy from NHLBI for her tireless effort regarding the Lung-HIV study, in particular, and lung disease in the HIV-infected population, in general.
1. Cohen BA, Pomeranz S, Rabinowitz JG, et al.. Pulmonary complications of AIDS: radiologic features. AJR Am J Roentgenol. 1984;143:115–122.
2. Kuhlman JE, Fishman EK, Hruban RH, et al.. Diseases of the chest in AIDS: CT diagnosis. Radiographics. 1989;9:827–857.
3. Murray JF, Mills J. Pulmonary infectious complications of human immunodeficiency virus infection. Part I. Am Rev Respir Dis. 1990;141:1356–1372.
4. Murray JF, Mills J. Pulmonary infectious complications of human immunodeficiency virus infection. Part II. Am Rev Respir Dis. 1990;141:1582–1598.
5. Braithwaite RS, Justice AC, Chang CC, et al.. Estimating the proportion of patients infected with HIV who will die of comorbid diseases. Am J Med. 2005;118:890–898.
6. Crum NF, Riffenburgh RH, Wegner S, et al.. Comparisons of causes of death and mortality rates among HIV-infected persons: analysis of the pre-, early, and late HAART (highly active antiretroviral therapy) eras. J Acquir Immune Defic Syndr. 2006;41:194–200.
7. Lau B, Gange SJ, Moore RD. Risk of non-AIDS-related mortality may exceed risk of AIDS-related mortality among individuals enrolling into care with CD4+ counts greater than 200 cells/mm3. J Acquir Immune Defic Syndr. 2007;44:179–187.
8. Marin B, Thiebaut R, Bucher HC, et al.. Non-AIDS-defining deaths and immunodeficiency in the era of combination antiretroviral therapy. AIDS. 2009;23:1743–1753.
9. Mocroft A, Brettle R, Kirk O, et al.. Changes in the cause of death among HIV positive subjects across Europe: results from the EuroSIDA study. AIDS. 2002;16:1663–1671.
10. Sider L, Gabriel H, Curry DR, et al.. Pattern recognition of the pulmonary manifestations of AIDS on CT scans. Radiographics. 1993;13:771–784; discussion 785–776.
11. McGuinness G, Scholes JV, Garay SM, et al.. Cytomegalovirus pneumonitis: spectrum of parenchymal CT findings with pathologic correlation in 21 AIDS patients. Radiology. 1994;192:451–459.
12. McGuinness G, Gruden JF, Bhalla M, et al.. AIDS-related airway disease. AJR Am J Roentgenol. 1997;168:67–77.
13. Logan PM, Finnegan MM. Pulmonary complications in AIDS: CT appearances. Clin Radiol. 1998;53:567–573.
14. Richards PJ, Armstrong P, Parkin JM, et al.. Chest imaging in AIDS. Clin Radiol. 1998;53:554–566.
15. Kuhlman JE. Imaging pulmonary disease in AIDS: state of the art. Eur Radiol. 1999;9:395–408.
16. Padley SP, King LJ. Computed tomography of the thorax in HIV disease. Eur Radiol. 1999;9:1556–1569.
17. Marchiori E, Muller NL, Soares Souza A Jr, et al.. Pulmonary disease in patients with AIDS: high-resolution CT and pathologic findings. AJR Am J Roentgenol. 2005;184:757–764.
18. Allen CM, Al-Jahdali HH, Irion KL, et al.. Imaging lung manifestations of HIV/AIDS. Ann Thorac Med. 2010;5:201–216.
19. Lichtenberger JP III, Sharma A, Zachary KC, et al.. What a differential a virus makes: a practical approach to thoracic imaging findings in the context of HIV infection–part 1, pulmonary findings. AJR Am J Roentgenol. 2012;198:1295–1304.
20. Lichtenberger JP III, Sharma A, Zachary KC, et al.. What a differential a virus makes: a practical approach to thoracic imaging findings in the context of HIV infection–part 2, extrapulmonary findings, chronic lung disease, and immune reconstitution syndrome. AJR Am J Roentgenol. 2012;198:1305–1312.
21. Kuhlman JE, Knowles MC, Fishman EK, et al.. Premature bullous pulmonary damage in AIDS: CT diagnosis. Radiology. 1989;173:23–26.
22. Diaz PT, Clanton TL, Pacht ER. Emphysema-like pulmonary disease associated with human immunodeficiency virus infection. Ann Intern Med. 1992;116:124–128.
23. Guillemi SA, Staples CA, Hogg JC, et al.. Unexpected lung lesions in high resolution computed tomography (HRCT) among patients with advanced HIV disease. Eur Respir J. 1996;9:33–36.
24. King MA, Neal DE, St John R, et al.. Bronchial dilatation in patients with HIV infection: CT assessment and correlation with pulmonary function tests and findings at bronchoalveolar lavage. AJR Am J Roentgenol. 1997;168:1535–1540.
25. Diaz PT, King MA, Pacht ER, et al.. Increased susceptibility to pulmonary emphysema among HIV-seropositive smokers. Ann Intern Med. 2000;132:369–372.
26. Diaz PT, Wewers MD, King M, et al.. Regional differences in emphysema scores and BAL glutathione levels in HIV-infected individuals. Chest. 2004;126:1439–1442.
27. Gingo MR, He J, Wittman C, et al.. Contributors to diffusion impairment in HIV-infected persons. Eur Respir J. 2014;43:195–203.
28. Gelman M, King MA, Neal DE, et al.. Focal air trapping in patients with HIV infection: CT evaluation and correlation with pulmonary function test results. AJR Am J Roentgenol. 1999;172:1033–1038.
29. Traill ZC, Miller RF, Shaw PJ. CT appearances of intrathoracic Kaposi's sarcoma in patients with AIDS. Br J Radiol. 1996;69:1104–1107.
30. Moskovic E, Miller R, Pearson M. High resolution computed tomography of Pneumocystis carinii pneumonia in AIDS. Clin Radiol. 1990;42:239–243.
31. Cattamanchi A, Nahid P, Marras TK, et al.. Detailed analysis of the radiographic presentation of Mycobacterium kansasii lung disease in patients with HIV infection. Chest. 2008;133:875–880.
32. Hartman TE, Primack SL, Muller NL, et al.. Diagnosis of thoracic complications in AIDS: accuracy of CT. AJR Am J Roentgenol. 1994;162:547–553.
33. Gruden JF, Huang L, Webb WR, et al.. AIDS-related Kaposi sarcoma of the lung: radiographic findings and staging system with bronchoscopic correlation. Radiology. 1995;195:545–552.
34. McGuinness G, Scholes JV, Jagirdar JS, et al.. Unusual lymphoproliferative disorders in nine adults with HIV or AIDS: CT and pathologic findings. Radiology. 1995;197:59–65.
35. Staples CA, Kang EY, Wright JL, et al.. Invasive pulmonary aspergillosis in AIDS: radiographic, CT, and pathologic findings. Radiology. 1995;196:409–414.
36. Edinburgh KJ, Jasmer RM, Huang L, et al.. Multiple pulmonary nodules in AIDS: usefulness of CT in distinguishing among potential causes. Radiology. 2000;214:427–432.
37. Jasmer RM, Edinburgh KJ, Thompson A, et al.. Clinical and radiographic predictors of the etiology of pulmonary nodules in HIV-infected patients. Chest. 2000;117:1023–1030.
38. Gold JA, Rom WN, Harkin TJ. Significance of abnormal chest radiograph findings in patients with HIV-1 infection without respiratory symptoms. Chest. 2002;121:1472–1477.
39. Richards PJ, Riddell L, Reznek RH, et al.. High resolution computed tomography in HIV patients with suspected Pneumocystis carinii pneumonia and a normal chest radiograph. Clin Radiol. 1996;51:689–693.
40. Hidalgo A, Falco V, Mauleon S, et al.. Accuracy of high-resolution CT in distinguishing between Pneumocystis carinii pneumonia and non- Pneumocystis carinii pneumonia in AIDS patients. Eur Radiol. 2003;13:1179–1184.
41. McGuinness G, Naidich DP, Garay S, et al.. AIDS associated bronchiectasis: CT features. J Comput Assist Tomogr. 1993;17:260–266.
42. Crothers K, Huang L, Goulet JL, et al.. HIV infection and risk for incident pulmonary diseases in the combination antiretroviral therapy era. Am J Respir Crit Care Med. 2011;183:388–395.
43. Hull MW, Phillips P, Montaner JS. Changing global epidemiology of pulmonary manifestations of HIV/AIDS. Chest. 2008;134:1287–1298.
44. Madeddu G, Fois AG, Calia GM, et al.. Chronic obstructive pulmonary disease: an emerging comorbidity in HIV-infected patients in the HAART era? Infection. 2013;41:347–353.
45. Shiels MS, Cole SR, Kirk GD, et al.. A meta-analysis of the incidence of non-AIDS cancers in HIV-infected individuals. J Acquir Immune Defic Syndr. 2009;52:611–622.
46. Guaraldi G, Besutti G, Scaglioni R, et al.. The burden of image based emphysema and bronchiolitis in HIV-infected individuals on antiretroviral therapy. PLoS One. 2014;9:e109027.
47. Drummond MB, Kirk GD, Astemborski J, et al.. Association between obstructive lung disease and markers of HIV infection in a high-risk cohort. Thorax. 2012;67:309–314.
48. Attia EF, Akgun KM, Wongtrakool C, et al.. Increased risk of radiographic emphysema in HIV is associated with elevated soluble CD14 and nadir CD4. Chest. 2014;146:1543–1553.
49. Sigel K, Wisnivesky J, Shahrir S, et al.. Findings in asymptomatic HIV-infected patients undergoing chest computed tomography testing: implications for lung cancer screening. AIDS. 2014;28:1007–1014.
50. Busi Rizzi E, Schinina V, Palmieri F, et al.. Radiological patterns in HIV-associated pulmonary tuberculosis: comparison between HAART-treated and non-HAART-treated patients. Clin Radiol. 2003;58:469–473.
51. Marras TK, Morris A, Gonzalez LC, et al.. Mortality prediction in pulmonary Mycobacterium kansasii infection and human immunodeficiency virus. Am J Respir Crit Care Med. 2004;170:793–798.
52. Samperiz G, Guerrero D, Lopez M, et al.. Prevalence of and risk factors for pulmonary abnormalities in HIV-infected patients treated with antiretroviral therapy. HIV Med. 2014;15:321–329.
53. Crothers K, Thompson BW, Burkhardt K, et al.. HIV-associated lung infections and complications in the era of combination antiretroviral therapy. Proc Am Thorac Soc. 2011;8:275–281.
54. Crothers K, McGinnis K, Kleerup E, et al.. HIV infection is associated with reduced pulmonary diffusing capacity. J Acquir Immune Defic Syndr. 2013;64:271–278.
55. Miller MR, Hankinson J, Brusasco V, et al.. Standardisation of spirometry. Eur Respir J. 2005;26:319–338.
56. Macintyre N, Crapo RO, Viegi G, et al.. Standardisation of the single-breath determination of carbon monoxide uptake in the lung. Eur Respir J. 2005;26:720–735.
57. Hankinson JL, Odencrantz JR, Fedan KB. Spirometric reference values from a sample of the general U.S. population. Am J Respir Crit Care Med. 1999;159:179–187.
58. Neas LM, Schwartz J. The determinants of pulmonary diffusing capacity in a national sample of U.S. adults. Am J Respir Crit Care Med. 1996;153:656–664.
59. Leader JK, Zheng B, Rogers RM, et al.. Automated lung segmentation in X-ray computed tomography: development and evaluation of a heuristic threshold-based scheme. Acad Radiol. 2003;10:1224–1236.
60. Muller NL, Staples CA, Miller RR, et al.. “Density mask”. An objective method to quantitate emphysema using computed tomography. Chest. 1988;94:782–787.
61. Gevenois PA, de Maertelaer V, De Vuyst P, et al.. Comparison of computed density and macroscopic morphometry in pulmonary emphysema. Am J Respir Crit Care Med. 1995;152:653–657.
62. Wang Z, Gu S, Leader JK, et al.. Optimal threshold in CT quantification of emphysema. Eur Radiol. 2013;23:975–984.
63. Lederer DJ, Enright PL, Kawut SM, et al.. Cigarette smoking is associated with subclinical parenchymal lung disease: the Multi-Ethnic Study of Atherosclerosis (MESA)-lung study. Am J Respir Crit Care Med. 2009;180:407–414.
64. Washko GR, Hunninghake GM, Fernandez IE, et al.. Lung volumes and emphysema in smokers with interstitial lung abnormalities. N Engl J Med. 2011;364:897–906.
65. Podolanczuk A, Sonti R, Kawut SM, et al.. Cross-sectional associations of ICAM-1 and HDL with subclinical interstitial lung disease (ILD): the MESA Lung Fibrosis Study. Am J Respir Crit Care Med. 2014;189:A3928.
66. Duffy P, Wang X, Lin PH, et al.. HIV Nef protein causes endothelial dysfunction in porcine pulmonary arteries and human pulmonary artery endothelial cells. J Surg Res. 2009;156:257–264.
67. Kline ER, Sutliff RL. The roles of HIV-1 proteins and antiretroviral drug therapy in HIV-1-associated endothelial dysfunction. J Investig Med. 2008;56:752–769.
68. Bankier AA, De Maertelaer V, Keyzer C, et al.. Pulmonary emphysema: subjective visual grading versus objective quantification with macroscopic morphometry and thin-section CT densitometry. Radiology. 1999;211:851–858.
69. Cavigli E, Camiciottoli G, Diciotti S, et al.. Whole-lung densitometry versus visual assessment of emphysema. Eur Radiol. 2009;19:1686–1692.
70. Cederlund K, Bergstrand L, Hogberg S, et al.. Visual classification of emphysema heterogeneity compared with objective measurements: HRCT vs spiral CT in candidates for lung volume reduction surgery. Eur Radiol. 2002;12:1045–1051.
71. COPDGene CT Workshop Group, Barr RG, Berkowitz EA, et al.. A combined pulmonary-radiology workshop for visual evaluation of COPD: study design, chest CT findings and concordance with quantitative evaluation. COPD. 2012;9:151–159.
72. Hersh CP, Washko GR, Jacobson FL, et al.. Interobserver variability in the determination of upper lobe-predominant emphysema. Chest. 2007;131:424–431.
73. Mets OM, Smit EJ, Mohamed Hoesein FA, et al.. Visual versus automated evaluation of chest computed tomography for the presence of chronic obstructive pulmonary disease. PLoS One. 2012;7:e42227.
74. Hersh CP, Make BJ, Lynch DA, et al.. Non-emphysematous chronic obstructive pulmonary disease is associated with diabetes mellitus. BMC Pulm Med. 2014;14:164.
75. Johannessen A, Skorge TD, Bottai M, et al.. Mortality by level of emphysema and airway wall thickness. Am J Respir Crit Care Med. 2013;187:602–608.
76. Yahaba M, Kawata N, Iesato K, et al.. The effects of emphysema on airway disease: correlations between multi-detector CT and pulmonary function tests in smokers. Eur J Radiol. 2014;83:1022–1028.