Association of Chronic Obstructive Pulmonary Disease Severity and Pneumocystis
Alison Morris MD, MS1,2*, Frank C. Sciurba MD2, Irina P. Lebedeva3, Andrew Githaiga
MD2, W. Mark Elliott PhD4, James C. Hogg MD4, Laurence Huang MD5, Karen A.
1Department of Medicine, Division of Pulmonary and Critical Care Medicine, University
of Southern California, Los Angeles, CA
2Department of Medicine, Division of Pulmonary, Allergy, and Critical Care Medicine,
University of Pittsburgh, Pittsburgh, PA
3Department of Immunology, University of Pittsburgh, Pittsburgh, PA
4Universityof British Columbia, McDonald Research Laboratories, St. Paul’s Hospital,
Vancouver, BC, Canada
5Department of Medicine, San Francisco General Hospital, University of California, San
*Corresponding author: Alison Morris, MD, MS
Division of Pulmonary and Critical Care Medicine
2011 Zonal Avenue, HMR 911
Los Angeles, CA 90033
Supported by NIH HL072837 (AM), HL072117 (LH), and Canadian Institute for Health
Research #7246 (WME, JCH).
This manuscript has an Online Data Supplement which is accessible through this issue's
table of contents online at www.atsjournals.org.
Running head: COPD severity and Pneumocystis
Subject category: 50
Word count: 3221
AJRCCM Articles in Press. Published on April 29, 2004 as doi:10.1164/rccm.200401-094OC
Copyright (C) 2004 by the American Thoracic Society.
Factors modulating the variable progression of chronic obstructive pulmonary disease
(COPD) are largely unknown, but infectious agents may play a role. Because
Pneumocystis has previously been shown to induce a CD8+ lymphocyte- and neutrophil-
predominant response similar to that in COPD, we explored the association of the
organism with accelerated disease progression. We examined Pneumocystis colonization
rates in lung tissue obtained during lung resection or transplantation in smokers with a
range of airway obstruction severity and in a control group with lung diseases other than
COPD. Using nested polymerase chain reaction, Pneumocystis colonization was detected
in 36.7% of patients with very severe COPD (Global Health Initiative on Obstructive
Lung Disease [GOLD] stage IV) compared to 5.3% of smokers with normal lung
function or less severe COPD (stages 0, I, II, and III)(p=0.004) and to 9.1% of controls
(p=0.007). Colonized subjects exhibited more severe airway obstruction (median
FEV1=21% predicted vs. 62% in non-colonized, p=0.006). GOLD IV was the strongest
predictor of Pneumocystis colonization (odds ratio=7.3, 95% confidence interval=2.4-
22.4, p<0.001) and was independent of smoking history. We conclude that there is a
strong association between Pneumocystis colonization and severity of airflow obstruction
in smokers, suggesting a possible pathogenic link with COPD progression.
Abstract word count: 199
Key words: Pneumocystis jiroveci, chronic obstructive pulmonary disease, epidemiology
Smoking has long been recognized as the primary risk factor for the development
of chronic obstructive pulmonary disease (COPD), but factors that determine which
smokers will develop significant disease are largely unknown. Recent interest has
focused on the potential role of infectious agents such as adenovirus, Chlamydia
pneumoniae, and other bacteria as co-factors in accelerating the progression of airway
obstruction (1-5). Pneumocystis jiroveci (formerly Pneumocystis carinii f. sp. hominis)
(6) is an eukaryotic opportunistic pathogen that causes pneumonia in
immunocompromised individuals and may be another pathogen involved in the
progression of COPD.
Although non-immunosuppressed hosts rarely develop Pneumocystis pneumonia
(PCP), use of the polymerase chain reaction (PCR) has demonstrated that some groups of
subjects have low levels of Pneumocystis DNA present in their lungs (7-11).
Pneumocystis in these cases, which likely represents colonization or asymptomatic
carriage, may lead to an exaggerated lung inflammatory response consisting primarily of
CD8+ lymphocytes and neutrophils (12-16). These same cell types are thought to be
important in the pathogenesis of COPD, and their numbers in the lung correlate with
severity of airflow obstruction (2, 17-21).
Previous data are inconclusive in linking Pneumocystis colonization with COPD.
One study found equivalent colonization rates in subjects with COPD compared to
subjects with other lung diseases (22). Other data suggest that Pneumocystis colonization
may be increased among those with COPD, but these studies were based on small
numbers of subjects, did not document COPD by pulmonary function testing or
pathology, and did not control for factors that might influence colonization or severity of
COPD such as smoking (9, 11). We conducted a cross-sectional analysis to determine if
Pneumocystis colonization is associated with severity of COPD independent of smoking
history. Some of the results of this study have been previously reported in the form of an
abstract (23, 24).
Word count: 521
Subjects: The smokers group were current or former smokers categorized according to
the Global Health Initiative on Obstructive Lung Diseases (GOLD) classification (25).
GOLD 0 (normal spirometry, at risk), I (mild COPD), II (moderate COPD), and III
(severe COPD) subjects were undergoing lung resection. GOLD IV (very severe COPD)
subjects were undergoing lung transplantation. Control subjects were undergoing lung
transplantation for primary pulmonary parenchymal disorders other than COPD. Lung
tissue was obtained during surgery and snap-frozen. The University of Pittsburgh and
University of British Columbia Institutional Review Boards approved the protocols.
DNA preparation and PCR amplification: DNA was extracted from a 1 cm3 sample of
lung tissue, and nested PCR performed at the Pneumocystis mitochondrial large sub-unit
(mtLSU) ribosomal RNA gene as previously described (26). Negative and positive
(DNA from lung tissue known to contain human Pneumocystis) controls were included.
All PCR was performed by personnel blinded to the subject identities, and all
reactions were carried out in an identical manner.
DNA sequencing: PCR products were purified and sequenced as previously described
and determined to be Pneumocystis jiroveci (27).
Data collection: Clinical data for GOLD IV and control patients were obtained from a
prospective database. Demographic information included age, gender, and race. Primary
diagnoses resulting in transplantation were determined. Subjects were defined as
undergoing transplantation for COPD if they carried a diagnosis of emphysema/COPD or
alpha-one antitrypsin deficiency based on transplant pulmonologists’ evaluations. We
included those with alpha-one antitrypsin deficiency because factors mediating the
variability of airway obstruction in these patients are poorly understood, even in
individuals homozygous for the condition. Results were not significantly changed when
excluding those with alpha-one antitrypsin deficiency. All other diagnoses were
considered non-COPD-related. Other information recorded included spirometry,
supplemental oxygen use, and diabetes mellitus. Smoking history included having ever
smoked and number of pack years smoked; subjects were required to have discontinued
smoking for at least six months prior to transplantation. Although respiratory cultures
from native lungs were not routinely tested, subjects were free of overt clinical infections
at transplantation. Type and dose of pre-transplant immunosuppression were recorded, as
was use of trimethoprim-sulfamethoxazole. Clinical data for smokers in GOLD 0-III
were obtained from medical records review and included age, gender, and smoking
Statistical analysis: Stata 7 (Stata Corporation, College Park, TX) was used for analysis,
and significance determined for a p-value< 0.05. Colonization status was determined for
the smokers as a group and by GOLD stage. Rates of colonization were compared
among GOLD stages using test of trend and Fisher’s exact. Odds ratios for colonization
were computed by comparing stage IV to other stages combined and to controls.
Severity of obstruction based on FEV1/FVC and FEV1 and colonization was assessed
using Mann-Whitney rank sum. Clinical characteristics of GOLD IV subjects were
compared to controls. Univariate analyses were performed to determine clinical
predictors of colonization in the entire cohort. For continuous variables, either Mann-
Whitney rank sum or t-tests were used. Chi-square or Fisher’s exact were used for
categorical variables. Spirometric measurements were expressed as percent predicted of
normal values (28). Dichotomous variables using values above and below the median
were created for analysis of spirometry values.
A total of 68 smokers were studied (10 in GOLD 0, 10 in GOLD I, 10 in GOLD
II, 8 in GOLD III, 30 in GOLD IV). Most were male (73.1%) and median age was 60
years (range 41-84). Median pack years of smoking were 60 (range 9 to 172) and median
FEV1 as percent of predicted was 46% (range 12-120%). Subject characteristics for each
GOLD stage are shown in Table 1. A total of 13 smokers (19.1%) were colonized with
Pneumocystis as determined by a positive nested PCR result and confirmed by DNA
Pneumocystis colonization and severity of COPD
Among smokers, Pneumocystis colonization was associated with a higher GOLD
stage (Table 2). Colonized subjects appeared similar to non-colonized subjects in terms
of age, gender, and pack year smoking history. Colonized subjects had a higher median
GOLD stage than non-colonized subjects (4 vs. 2, p=0.001). No subjects in GOLD stage
0 or I were colonized. Furthermore, the rate of colonization increased with increasing
GOLD stage (p=0.002 for test of trend)(Figure 1), and the odds ratio for colonization for
stage IV compared to the other stages was 10.4 (95% CI=2.1-51.9, p=0.004). In order to
determine if Pneumocystis colonization increased COPD risk independent of smoking
history, we adjusted the odds ratio for pack year history and found that the association
with COPD severity as determined by GOLD stage actually became stronger (adjusted
odds ratio=21, 95%CI=2.3-191.6, p=0.007). Average pack year history by stage did not
differ significantly (p=0.39)(Table 1).
Colonized subjects had significantly worse airway obstruction than non-colonized
subjects (Table 2). Median percent predicted FEV1 for colonized subjects was 21%
compared to 62% for non-colonized (p=0.006), and the ratio of FEV1/FVC was similarly
decreased (32% in colonized vs. 54% in non-colonized, p=0.001). Colonized subjects
were significantly more likely to have FEV1 and FEV1/FVC values below the cohort
median, even when adjusted for pack year smoking history (adjusted OR=16.7,
95%CI=1.78-157.2, p=0.014 for FEV1; adjusted OR=14.4, 95%CI=1.6-127.7, p=0.016
Comparison to controls
Because colonization in the patients with severe COPD might have resulted from
factors associated with end-stage lung disease rather than COPD itself, we compared
clinical characteristics of the GOLD IV subjects and controls (Table 3). As control
subjects were all undergoing lung transplantation, they had severe lung diseases and were
similar in overall health status to the GOLD IV subjects.
There were 44 control subjects who underwent lung transplantation for diagnoses
other than COPD. Reasons for transplantation included cystic fibrosis (n=15) and
idiopathic pulmonary fibrosis (n=12). No subject had a history of immunodeficiency or
PCP. Spirometry was performed an average of 194 days prior to transplantation. GOLD
IV subjects were significantly older than controls (median age=58 years vs. 42 years,
p<0.001). They were similar in terms of gender, race, oxygen dependence, and use of
trimethoprim-sulfamethoxazole at time of transplantation. More control subjects tended
to be receiving prednisone and at a higher dose, although the difference was not
statistically significant. Only control patients carried a diagnosis of diabetes mellitus
(15.9% vs. 0%, p=0.04); however, diabetes was not related to colonization (p=0.73). As
would be expected, GOLD IV subjects were more likely to have smoked and had a
greater pack year history of smoking. They also had a significantly lower FEV1 percent
predicted (21% vs. 31%, p=0.004) and FEV1/FVC percent (32% vs. 69%, p<0.001).
Four control subjects (9.1%) were colonized with Pneumocystis in the explanted
lung at time of transplantation. When compared to controls, GOLD IV subjects had a
significantly higher rate of colonization (36.7%, OR=5.8, 95%CI=1.6-20.6, p=0.007).
The rate of colonization was similar in controls and other GOLD stages (5.4% for GOLD
0-III, OR=0.55, 95%CI=0.1-3.2, p=0.56).
Two subjects developed PCP after transplantation. Both were GOLD IV subjects
who had Pneumocystis colonization of their native lungs before transplantation. These
cases occurred between 104 and 154 days after transplantation and were diagnosed by
bronchoscopic alveolar lavage. The subjects both had single lung transplantation and
were receiving prophylaxis with trimethoprim-sulfamethoxazole at time of diagnosis.
Predictors of Pneumocystis colonization
Univariate analyses were performed to determine clinical predictors of
colonization in the entire cohort of smokers and controls (Table 4). Data for some
variables were not available for the GOLD 0-III subjects; however, none of these
variables (prednisone use, diabetes mellitus, trimethoprim-sulfamethoxazole use, oxygen
use) was found to be relevant to colonization when comparing GOLD IV to controls.
Univariate analyses demonstrated that age, gender, history of smoking, and number of
pack-years smoked were not related to colonization risk in the cohort as a whole. The
most significant clinical predictor of Pneumocystis colonization was a diagnosis of very
severe (GOLD IV) COPD (OR=7.3, 95% CI=2.4-22.4, p<0.001). In addition to COPD,
both a low FEV1 and a low FEV1/FVC were predictive of colonization. FEV1 and
FEV1/FVC were clinically and statistically correlated with a diagnosis of severe COPD
(p<0.001), and once Pneumocystis colonization risk was adjusted for a severe COPD
diagnosis, neither were independently predictive of risk.
In this study, we demonstrated that Pneumocystis colonization was associated
with severity of airway obstruction in a cohort of patients at risk for COPD. Colonization
was not increased in subjects with severe lung disease from causes other than COPD. In
addition, the association of colonization with severe COPD was not explained by clinical
factors such as age, use of immunosuppressive medication, or the presence of co-morbid
Use of PCR for detection of Pneumocystis has led to discovery of Pneumocystis
DNA in subjects without clinical PCP. Presence of Pneumocystis in respiratory
specimens from subjects without signs or symptoms of clinical infection and without
progression to PCP has been defined as colonization. Colonization has not generally
been detected in healthy subjects (29-31), but rates of colonization in HIV-infected
subjects may be as high as 69% (32). Rates of colonization in patients with chronic lung
diseases range from 7 to 41% (7-10), and colonization in the non-HIV-infected host has
been shown to be more common in subjects with a CD4+ Tcell count below 400cells/ul
and aCD4+ T cell/CD8+ Tcell ratio less than one (33).
This study examined clinical variables that might affect Pneumocystis
colonization in a well-characterized subject group. By comparing colonization in
subjects with equivalent smoking histories, we were able to assess Pneumocystis as a
factor that may affect differential progression of COPD in smokers. By comparing
colonization status in transplant patients with equivalent severity of lung disease, we
were able to examine the specific association of Pneumocystis and COPD. We have thus
shown an association between Pneumocystis colonization and severity of COPD and
demonstrated that colonization is not a result of other clinical factors or co-morbid
Previous work has been contradictory regarding detection of Pneumocystis in
those with COPD. A recent study found that there was no increase in the risk of
colonization in subjects with COPD (22). These results may differ from ours because the
authors studied a population with less severe obstruction. Average FEV1 in their study
was 2.1 L/min, a value much higher than the average of our colonized COPD patients
(mean FEV1=0.8 L/min). Another series that studied patients with chronic sputum
production during a time of exacerbation found that 10% had Pneumocystis detectable by
standard staining methods of sputum (7). This rate might be lower than in our population
because Pneumocystis colonization was determined by microscopic examination and not
by PCR. Probst and colleagues reported a colonization rate of 41% in 37 subjects with
COPD based on nested PCR of various respiratory samples (9). Helweg-Larsen et al.
recently found that 43% of 23 subjects with COPD were colonized (11). However, these
subjects were hospitalized with respiratory symptoms and therefore might have a
different colonization rate than at baseline. Pulmonary function data and smoking history
were not reported for subjects in either study, and results were not controlled for other
clinical variables that might influence colonization. An additional difficulty with all
studies of Pneumocystis colonization is the inability of PCR to determine viability of the
Other microbiologic agents have also been implicated in the pathogenesis of
COPD progression in smokers. Retamales and colleagues, using a study design similar to
ours, have recently reported a higher rate of detection of adenovirus E1A protein in lung
tissue of smokers with emphysema compared to age- and smoking-matched controls
without evidence of airway obstruction (2). Based on elevated levels of CD8+
lymphocytes and interferon-γ in lung samples, the authors speculated that the presence of
latent adenovirus in the lungs provokes a heightened inflammatory response that leads to
worsening of airway obstruction. Data in animals support this hypothesis, as guinea pigs
with adenoviral infections have a more pronounced inflammatory response and
accelerated emphysema development compared to non-infected controls (34).
Occult bacterial infections with organisms such as Haemophilus influenzae,
Streptococcus pneumoniae, and Moraxella catarrhalis have also been implicated in
COPD progression. The vicious circle hypothesis proposed by Sethi and Murphy
postulates that bacterial colonization of the lower respiratory tract leads to amplification
of chronic inflammation and worsening airway obstruction (3, 35). Previous work has
shown high levels of bacterial colonization in subjects with COPD and also demonstrated
that bacterial colonization is associated with increased numbers of inflammatory cells and
cytokines (5, 36). A recent study found that subjects with increases in the sputum
bacterial load had higher sputum interleukin (IL)-8 levels and more rapid declines in
FEV1 than those with lower levels of bacterial colonization (5). Other investigators have
demonstrated elevated IL-8, tumor necrosis factor-α, and neutrophil elastase in subjects
with bacterial colonization during acute exacerbations (17, 36, 37).
Pneumocystis may accelerate progression to COPD through a similar
inflammatory mechanism. A recent study of simian immunodeficiency virus (SIV)-
infected macaques reported that animals colonized with Pneumocystis had a T-
lymphocyte influx in lung lavage fluid comprised of over 90% CD8+ T cells (16). This
influx was greater than that seen in SIV-infected animals without Pneumocystis and was
not reflected by peripheral blood lymphocyte levels. Lung neutrophil levels, IL-8, and
interferon-γ were also significantly increased in Pneumocystis-colonized macaques (26).
This inflammatory response developed well before the onset of PCP in the animals.
Human studies of HIV- and non-HIV-infected subjects with PCP have also shown a
sharp increase in the number of CD8+ cells and neutrophils in the lung (12-15).
The clinical significance of inflammation resulting from Pneumocystis is
unknown, but the host response in Pneumocystis resembles that in COPD. COPD is
characterized by inflammation of peripheral airways and destruction of lung parenchyma
resulting in airflow limitation. The exact nature of the inflammatory changes that occur
in COPD and how these differ from those seen in smokers without lung disease are not
completely understood, but high levels of neutrophils and lymphocytes, particularly
CD8+ lymphocytes, appear to play a critical role. The number of CD8+ lymphocytes in
the lung correlates directly with the degree of airflow limitation (2, 18-20). Numbers of
neutrophils and IL-8 levels are also correlated with severity of disease (17, 19, 21).
There is some previous evidence linking Pneumocystis to COPD-like changes.
The Pulmonary Complications of HIV Infection Study, a prospective cohort of over
1,100 HIV-infected subjects followed for a median of four years, obtained serial
pulmonary function studies in subjects at pre-defined intervals and after an episode of
pneumonia. HIV-infected subjects with PCP had accelerated declines in FEV1,
FEV1/FVC,and diffusing capacity beyond that expected from age and smoking history.
These changes in pulmonary function were indistinguishable from those seen clinically
with COPD and persisted for years after the acute infection resolved (38). Furthermore,
the accelerated progression of emphysema in HIV-infected subjects described by Diaz
and co-workers may be present in the absence of an overt history of PCP infection and
may be related to occult infection with microorganisms such as Pneumocystis (39-42).
Specific factors that render the lung susceptible to Pneumocystis colonization are
not known. Previous work has found that smoking is an independent risk factor for
colonization in HIV-infected subjects (43), but the mechanism through which this occurs
has not been elucidated. It is possible that structural remodeling associated with COPD
renders smokers more likely to become colonized and/or less able to clear subclinical
infection. Once colonization with Pneumocystis is established, the organism might
stimulate chronic inflammation that facilitates accelerated decline with or without the
persistence of tobacco exposure. Possible factors that may increase the likelihood of
colonization or decrease the lung’s ability to clear the organism include defects in
mucociliary clearance and surfactant abnormalities produced by smoking (44-48).
Although the association of Pneumocystis with a diagnosis of COPD is intriguing,
the current study cannot provide definitive conclusions regarding cause and effect.
Because subjects were tested at a single time point, we do not have information about the
time course of colonization relative to disease progression. We also cannot determine
whether Pneumocystis colonization results in acceleration of COPD or if some factor
unique to COPD results in Pneumocystis colonization. The design of the study, by using
a control group with similarly severe lung disease and overall health status, minimizes the
possibility that colonization is merely associated with end-stage lung disease. Because
we tested only for Pneumocystis, other occult infections may also have been present in
these subjects. It is interesting, however, that subjects with cystic fibrosis and
bronchiectasis, diseases in which airway colonization is prominent, did not have an
increased detection of Pneumocystis colonization. Pneumocystis may act alone or in
concert with other organisms to provoke inflammation and lung damage, and the
relationship of Pneumocystis and colonization with various organisms should be explored
in future studies. In addition to direct studies of colonization, studies of serologic
responses to Pneumocystis and other organisms might also help in understanding the role
of these infections in airway obstruction.
In summary, we have shown that there is a strong association of Pneumocystis
colonization and severity of airflow obstruction in subjects at risk for COPD.
Colonization is disproportionately increased among those with severe COPD compared to
15 Download full-text
other smokers and compared to subjects with end-stage lung diseases not related to
COPD. These findings suggest that Pneumocystis is an infectious agent that may play a
role in the accelerated progression of airway obstruction. The pronounced pulmonary
inflammatory response that occurs in response to Pneumocystis may contribute to the
pathogenesis of airway and parenchymal damage in smokers or individuals with
structural airway obstruction. Because infectious agents such as Pneumocystis are
potentially treatable, future studies are needed to further define the nature of
Pneumocystis colonization in COPD and its role in disease pathogenesis.