Abstract
Background and objective: Bactericidal/permeability-increasing protein (BPI) is a member of the pattern recognition receptors of the innate immune system. Recently, an association between genetic polymorphism in the BPI gene and a risk of airflow decline after transplantation was demonstrated, but whether these findings are reproducible in nontransplantation populations, such as those with COPD, is still unknown. The aim of this study is to explore the role of BPI in COPD. Methods: The genotypes of 107 patients with COPD and 110 control subjects were evaluated by polymerase chain reaction and polymorphism analysis of the BPI genes and ELISA analysis of the plasma BPI level. All subjects were men over 40 years old who smoked. Results: BPI mutation PstI (T→C) polymorphism in intron 5 was associated with an increased risk of developing COPD (OR 3.73, 95%CI: 1.62–9.10), and the frequency was significantly increased in the COPD group compared with the control group (26/107 [24.3%] vs 12/110 [10.9%], p = 0.002). In addition, COPD patients exhibited a decreased plasma level of BPI compared with the control group (10.6 ± 2.2 vs 23.4 ± 2.1ng/ml, p < 0.0001). Conclusions: BPI mutation (PstI in intron 5) and a decreased plasma BPI level were significant risk factors in susceptibility to COPD. These results demonstrate that BPI genetic mutation and impaired BPI production or release may result in airflow obstruction in smokers.
Introduction
Chronic obstructive pulmonary disease (COPD) is currently the fourth leading cause of death worldwide and the World Health Organization estimates it will be the third leading cause of death by 2020 (Citation1). The burden of COPD in Asia is currently greater than that in developed Western countries; the number of COPD patients in Asia exceeds by three times the total number of COPD patients in the rest of the world (Citation2).
The prevalence and mortality of COPD are rapidly increasing in Asian countries; this is mainly because of increasing exposure to environmental cigarette smoke and indoor and outdoor air pollution (Citation2, 3). Lipopolysaccharide (LPS) is one of the most potent stimulators of the innate immune system and is also one of the agents present in cigarette smoke and air pollution (Citation4, 5). Bioactive LPS has been reported to be present at high levels in cigarette smoke and is a major determinant of lung function decline (Citation6).
Chronic exposure to significant levels of LPS was found to be associated with the development of progressive irreversible airway obstruction, and results in persistent chronic pulmonary inflammation. Previous studies have found that lipopolysaccharide-induced inflammatory signaling is increased in COPD (emphysema) patients and may be involved in elastin destruction by causing increased release of elastolytic enzymes such as matrix metalloproteinase (Citation7, 8).
In addition, mice having long-term LPS exposure show chronic and persistent pulmonary inflammation, and the inflammatory and pathologic changes mimic changes observed in COPD patients with lung inflammation (Citation9, 10, Citation11). These studies indicate that LPS is an important pathogenic substance in cigarette smoke and air pollution contributing to the COPD that develops in susceptible cigarette smokers.
Bactericidal/permeability-increasing protein (BPI) is a 55-kDa protein found in the primary granules of human neutrophils, and has also been detected on the surface of intestinal and oral epithelial cells (Citation12, 13). More recently, BPI has also been detected in human small airways which constitutively express the BPI gene and produce the protein (Citation14). BPI is capable of inhibiting bacterial growth, activating bacterial phospholipid hydrolysis, and penetrating the inner bacterial membrane to dissipate electrochemical gradients required for bacterial viability. In addition, the high-affinity binding of human BPI to the lipid A region of LPS prevents the transfer of LPS to cellular receptors and facilitates its uptake by macrophages, leading to the elimination of LPS and LPS-induced inflammation (Citation12, 13, Citation15).
Recently, Chien et al. demonstrated for the first time that genetic polymorphism in the BPI gene significantly influences the risk of developing rapid airflow decline after transplantation and associated this with the phenotype of severe airflow obstruction. They hypothesized that a loss-of-function BPI mutation is likely the cause of rapid decline of lung function (Citation14). Whether these findings are reproducible in nontransplantation populations, such as patients with COPD, is unknown, but it is highly possible considering the close association among BPI, LPS and COPD.
The aim of this study is to explore the role of BPI in COPD. The genotypes of BPI in patients with COPD and control subjects will be evaluated by polymerase chain reaction (PCR) followed by restriction fragment length polymorphism (RFLP) analysis and their plasma BPI level will be evaluated by ELISA analysis. If we prove the hypothesis that BPI genetic variation and plasma level are associated with airflow obstruction in smokers, we might demonstrate the important role of BPI in the pathogenesis of COPD; this may present an opportunity for the development of novel therapies for BPI-deficient COPD patients with a rapid decline in lung function.
Patients and Methods
Study population
This study was a hospital-based case-control study. The subjects in this study consisted of 217 patients. The case group included 107 men 40 years old with smoking -related COPD (smoking history 10 pack-years), who were recruited from National Cheng-Kung University Hospital. They were diagnosed as having COPD on the basis of their medical history, chest radiographic findings, and spirometric results, according to Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2008 guidelines (Citation1). Inclusion criteria for COPD included the following: chronic airway symptoms and signs such as coughing, dyspnea, wheezing and chronic airway obstruction, defined as a ratio of post-bronchodilator forced expiratory volume in one second (FEV1)/forced vital capacity (FVC) of < 70%.
Patients were excluded if they had a history of asthma or lung cancer; patients with liver cirrhosis and acute exacerbation were also excluded from the study in order to avoid their interfering effects on serum BPI levels (Citation16, 17). The severity classifications according to GOLD stage (Citation1) for these 107 patients were mild (20/107; 18.7%), moderate (45/107; 42.1%), severe (29/107; 27.1%), and very severe (13/107; 12.1%).
The control group included 110 asymptomatic men 40 yrs old with a smoking history 10 pack-yrs without clinical symptoms of COPD and with normal pulmonary function (FEV1/FVC 70% and FEV1 80% predicted). Most of the subjects visited our hospital for health checkups. Subjects with liver cirrhosis, and bacterial infection were excluded to avoid their interfering effects on serum BPI levels. Only men were enrolled in this study, because in our population, very few women smoke or have COPD.
The study was approved by the Ethics Committee of National Cheng-Kung University Medical College and Hospital, and informed consents were obtained.
DNA extraction and polymerase chain reaction restriction fragment length polymorphism
Genomic DNA was isolated by Puregene commercial DNA isolation kits according to standard procedures (Gentra Systems, Inc, Minneapolis, MN, USA). The total volume of the PCR amplification mixture was 50 μL containing 100 ng of genomic DNA, 0.5 mM of a dNTP Mix, 0.5 U Taq DNA polymerase, 0.2 mM of each nucleotide, and 1.5 mM MgCl2. The amplification protocol was done using a Perkin Elmer DNA Thermal Cycler 9600.
Three polymorphisms in the BPI gene, a T to C polymorphism in intron 5, the silent nucleotide exchange G545C (Val182Val) and the polymorphism A645G resulting in the amino acid exchange Lys216Glu, were genotyped using PCR and RFLP analysis according to the method of Hubacek et al. (Citation18, 19). For T to C polymorphism in intron 5, the primer sequences were 5’-ATG GCT CAG AGA GGC TAA GTG GAG GAC AG-3’ (sense) and 5’-TAG CTC CTG GTG TGG TTG CCC ATT CC-3’ (antisense). For the restriction enzyme PstI, the size of the PCR products was 199 bp, and the sizes of the restriction fragments for the appropriate alleles were 199 bp for the T allele and 109 bp+ 90 bp for the C allele.
For the silent nucleotide exchange G545C (Val182Val), the primer sequences were 5’-TAA CCC CAC GTC AGG CAA GCC CAC C-3’ (sense) and 5’-AGC CAC AGT CCC GGG AGT CCA TAC TC-3’ (antisense). For the restriction enzyme TaqI, the size of the PCR products was 117 bp, and the sizes of the restriction fragments for the appropriate alleles were 117 bp for the G allele and 91 bp+ 26 bp for the C allele. For polymorphism A645G resulting in the amino acid exchange Lys216Glu, the primer sequences were 5’-ACT ATG GGA AGA CCT TAC TGA TTA C-3’ (sense) and 5’-CAG AGT CTG GAA ATA AGG TTG AAG C-3’ (antisense).
For the restriction enzyme HindIII, the size of the PCR products was 103 bp, and the sizes of the restriction fragments for the appropriate alleles were 103 bp for the G allele (Glu) and 79 bp+ 24 bp for the A (Lys) allele. The fragments were separated by electrophoresis through a 3% agarose gel, stained with ethidium bromide and train illuminated with ultraviolet light.
Plasma BPI level determinations
Blood was collected in tubes containing EDTA (Sherwood Medical, St. Louis, MO, USA); samples were put on ice immediately and kept on ice during the enteric preparation. Plasma was separated by centrifugation according to the method of Groenewegen et al. (Citation17). The supernatants were collected after centrifugation and stored at -70°C until analysis. Plasma BPI was measured by a sandwich enzyme-linked immunosorbent assay developed by Hycult Biotech (Hycult Biotechnology, Uden, the Netherlands).
Statistical analysis
Associations between specific genotypes and phenotypes were analyzed for significance by the two-tailed chi-square test and unpaired t-test for univariate analysis. A logistic regression model was used to calculate odds ratios (ORs) adjusted for age, smoking status and cumulative cigarette consumption in pack-years between COPD and control subjects. Multivariate analysis was used to assess the contributions of genetic factors, plasma level, age, and smoking (status and cumulative amount by pack-years) to the risk of developing of COPD in all smokers.
Analysis for the association between BPI level and COPD severity was adjusted by the patients’ medical history with chronic steroid and long-acting ß2-agonist use to avoid their interfering effect on the serum BPI level (Citation20).We used the JMP (Jump) software program (SAS Institute Inc., Cary, NC, USA) for multivariate analysis. Power analysis was calculated using QUANTO version 1.2.4 (www.hydra.usc.edu/GxE/) (2009). The genetic effects of the PstI (T→C) polymorphisms in intron 5 were assumed to be C allele dominant, because only heterozygous mutation was noted in the present study and the effects of the A645G (Lys216Glu) polymorphisms were assumed for both models for G allele dominant and recessive.
The population risk was assumed to be 10%. Adjustment for multiple corrections was made using the Bonferroni approach. In each analysis, a difference with a p value of < 0.05 was accepted as significant.
Results
The characteristics of the control and COPD subjects are summarized in . The influence ORs (95%CI) of age, smoking history (pack-years), and actual smoking status (current smoker) in the prevalence of COPD were 39.5 (10.8–158.5), 5.11 (0.69–46.5) and 0.70 (0.37–1.33), respectively. The mean age and smoking history were lower and the frequency of current smoking was higher in the control group than the COPD group; we adjusted for this incomplete matching in the analysis of the data. Polymorphisms for the BPI gene with silent nucleotide exchange G545C (Val182Val) were not found in our study. The Hardy-Weinberg equilibrium was used to test the genotypes for PstI (T→C) polymorphism in intron 5 and A645G (Lys216Glu) polymorphism in both the COPD and control groups, and no obvious deviations were found.
Table 1. Patient characteristic of COPD and control smokers
Subjects with PstI (T→C) polymorphism in intron 5 mutant were all heterozygotes with an increased risk of developing COPD (OR 3.73, 95%CI: 1.62–9.10). The frequency of this genotype was significantly higher in the COPD group than the control group (26/107 [24.3%] vs 12/110 [10.9%], p = 0.002) (). The power to detect genetic risks of 3.73 (OR) was 93% if the inheritance mode was C allele dominant. If the inheritance mode was C allele recessive, the power was 10%.
Table 2. Distribution of COPD associated with BPI genetic polymorphisms
Because only heterozygous mutation and no CC homozygous mutation were noted, the C allele recessive mode was not suitable in this study. For A645G polymorphism, the Glu/Glu genotype for BPI A645G (Lys216Glu) polymorphism was decreased in COPD patients compared with controls (OR: 0.52, 95%CI: 0.27–0.98; 65/107 [60.8%] vs 72/110 [65.5%], p = 0.045), but the p-values were not significant after Bonferroni correction for multiple tests.
The power was 64% to detect genetic risks of 0.52 (OR) if the inheritance mode was G allele recessive and the power was 48% to detect genetic risks of OR:0.32 (OR), 95%CI: 0.09–1.05, with a non-significant p-value, if the inheritance mode was G allele dominant. In the genetic models for the allele of Pst (T to C) in intron 5 mutation; the frequency of mutation allele (C allele) was higher in the COPD group than the control group (26/214, 12.2% vs 12/220, 5.5%), OR = 3.05, 95% CI: 1.40–6.96, p-value = 0.006.
In the genetic models for the allele of A645G polymorphism; the frequency of the G allele for A645G polymorphism was lower in the COPD group than the control group (163/214, 76.2% vs 177/220, 80.5%), OR = 0.54, 95% CI (0.32–0.90), p value = 0.018.
The frequency of the minor allele (C allele) of Pst (T to C) in intron 5 mutation was 5.5% (12/220) for the control group in the present study. The prevalence of the C allele of Pst (T to C) in intron 5 mutation (rs 6099106) has not been reported in the Single Nucleotide Polymorphisn database (dbSNP).
The prevalence of the minor allele of A645G polymorphism (A allele) (rs 4358188) in the dbSNP has been reported to range from 19.5% to 26.7% in Han Chinese populations (Asian). The frequency of the A allele in the present study was 19.6% (43/220) within the control group, consistent with that reported in the dbSNP.
COPD patients exhibited a decreased plasma level of BPI compared with the control group (10.6 ± 2.2 vs 23.4 ± 2.1ng/ml, p < 0.0001). The plasma BPI level was also associated with the severity of COPD according to the GOLD classification (1); patients in the very severe stage (FEV1 < 30% of *predicted) revealed a decreased BPI level compared with those in the mild, moderate and severe stages (FEV1 30% of predicted) (3.6 ± 4.5 vs 11.5 ± 1.7 ng/ml, p = 0.016) (). There was no significant correlation between plasma BPI level and BPI genotypes (Pstl in intron 5 and A645G polymorphism).
Table 3. Patient characteristic and plasma BPI level of COPD with very severe stage (FEV1 < 30% pred.) and mild to severe stages (FEV1 30% pred.)
Discussion
In the present study, a significant positive association between BPI mutation (Pstl in intron 5) and COPD was noted. Point mutations that occur in introns and affect the splice-site signals are considered to be splicing mutations. Mutations that affect sequences that are important for splicing modulation are likely to have a profound effect on the translated product. It has been estimated that at least 15% of point mutations that result in human genetic diseases cause RNA splicing defects (Citation21, 22). It is not known if BPI mutation (Pstl in intron 5) causes a splicing mutation which then results in the loss of function of BPI. However, it is possible, considering the results of our study.
The BPI A645G polymorphism that leads to the amino acid exchange Lys216Glu is suggested to be functionally effective and is associated with Crohn's disease (Citation23). Crohn's disease is a chronic inflammatory bowel disease. Environmental risk factors such as cigarette smoke and genetic factors such as alpha1-antitrypsin deficiency affect the pathogenesis of both COPD and Crohn's disease (Citation24). In addition, a large population-based study revealed that COPD patients had a significantly higher risk of Crohn's disease than control subjects without COPD in the general population (HR 2.72; 95%CI 2.33–3.18) (Citation25). These studies suggest that COPD and Crohn's disease may have common inflammatory pathways, including genetic variants of susceptibility for diseases. This hypothesis was consistent with our findings.
A previous study found a significant decrease in the frequency of the Glu/Glu genotype in Crohn's disease (Citation23). A protective effect similar to the disease susceptibility for COPD patients was noted in the present study (OR: 0.52, 95%CI: 0.27–0.98 for the Glu/Glu genotype; p-values were not significant after Bonferroni correction).
COPD is considered an inflammatory pulmonary disease with systemic inflammation manifestations; systemic inflammatory biomarkers are associated with disease severity (Citation26, 27, Citation28). Serum tumor necrosis factor-alpha (TNF-α) levels correlated with the severity of airways obstruction, osteoporosis and weight loss for COPD patients in one study (Citation26). Plasma BPI has been proven to inhibit LPS- mediated triggering of TNF-α release by monocytes in human study (Citation16).
Therefore, it is reasonable that decreased plasma BPI due to impaired production or release will cause increased LPS-mediated inflammation such as increased TNF-α, resulting in clinical manifestations in patients with COPD. Our finding was consistent with the hypothesis that COPD patients exhibited a significantly decreased plasma level of BPI compared with the control group; the plasma BPI level was also negatively associated with the severity of COPD.
One limitation of the present study was that only men were enrolled. The majority of patients with smoking-related COPD in our population are men (smoking prevalence for women is less than 5%) (Citation3).
In contrast to industrialized countries in the West, COPD morbidity remains male- predominant in Asian countries (Citation2). The second limitation was that this study did not validate in another independent cohort. The third limitation was the relatively small number of recruited subjects, limitations of the findings to the Chinese population. A large group of study subjects with varied characteristics should be considered in a future study to clearly demonstrate a significant contribution of these polymorphisms.
Another limitation was the BPI concentration in the epithelial mucosa was not evaluated in our study. The BPI concentration in the epithelial mucosa is a direct marker for COPD and the serum BPI concentration is an indirect marker. Evaluation of the BPI concentration in the epithelial mucosa would provide stronger evidence of the role of BPI in COPD than the serum BPI concentration.
In conclusion, in our population, BPI mutation (PstI in intron 5) and a decreased plasma BPI level were significant risk factors for susceptibility to smoking- related COPD. Treatment with a recombinant form of BPI is available (Citation29). If further studies can confirm that loss-of-function BPI genetic mutation and defective BPI production or release play major roles in airflow obstruction in patients with COPD, novel therapies can be developed to treat these patients. In addition, scientists can attempt some mechanism- based studies to verify the functional significance of these polymorphisms in context with various diseases in the future.
Declaration of Interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
Acknowledgments
This study was supported by grant NSC 95–2314-B-006–026 from the National Science Council and NCKUH-9804014 from National Cheng Kung University Hospital. We are grateful to Jia-Ling Wu and Sheng-Hsiang Lin for providing the statistical consulting services from the Biostatistics Consulting Center, National Cheng Kung University Hospital. All authors have no conflicts of interest related to the material in this manuscript.
References
- Global Initiative for Chronic Obstructive Lung Disease (GOLD). Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease, updated. Bethesda, MD: National Heart, Lung and Blood Institute/World Health Organization, 2008 (http://www.goldcopd.org).
- Tan WC, Ng TP. COPD in Asia: Where East meets West. Chest 2008; 133:517–527.
- Regional COPD Working Group. COPD prevalence in 12 Asia-Pacific countries and regions: Projections based on the COPD prevalence estimation model. Respirology 2003; 8:192–198.
- Vogelzang PFJ, van der Gulden JW, Folgering H, Kolk JJ, Heederik D, Preller L, Tielen MJM, van Schayck CP (1998) Endotoxin exposure as a major determinant of lung function decline in pig farmers. Am J Respire Crit Care Med 1998; 157:15–18.
- Thorne PS, Kulhankova K, Yin M, Cohn R, Arbes SJ, and Zeldin DC. Endotoxin exposure is a risk factor for asthma; the national survey of endotoxin in United States housing. Am J Respire Crit Care Med 2005; 172:1371–1377.
- Hasday JD, Bascom R, Costa JJ, Fitzgerald T, Dubin W. Bacterial endotoxin is an active component of cigarette smoke. Chest 1999; 115:829–835.
- Russell RE, Culpitt SV, Dematos C, Donnely L, Smith M, Wiggins J, Barnes PJ. Release and activity of matrix metalloproteinase-9 and tissue inhibitor of metalloprotease-1 by alveolar macrophages from patients with chronic obstructive pulmonary disease. Am J Respire Cell Mol Biol 2002; 26:602–609.
- Lim S, Roche N, Oliver BG, Mattos W, Barnes PJ, Chung KF. Balance of matrix metalloprotease-9 and tissue inhibitor of metalloprotease-1 from alveolar macrophages in cigarette smokers. Regulation by interleukin-10. Am J Respir Crit Care Med 2000; 162(pt1):1355–1360.
- Vernooy JH, Dentener MA, van Suylen RJ, Buurman WA, Wouters FM. Long-term intratracheal lipopolysaccharide exposure in mice results in chronic lung inflammation and persistent pathology. Am J Respir Cell Mol Biol 2002; 26:152–159.
- Sajjan U, Ganesan S, Comstock AT, Shim J, Wang Q, Nagarkar DR, Zhao Y, Doldsmith AM, Sonestein J, Linn MJ, Curtis JL, Hershenson MB. Elastase- and LPS- exposed mice display altered responses to rhinovirus infection. Am J Physiol Lung Cell Mol Physiol 2009; 297:L931–L944.
- Meng QR, Gideon KM, Harbo SJ, Renne RA, Lee MK. Gene expression profiling in lung tissue from mice exposed to cigarette smoke, lipopolysaccharide, or smoke plus lipopolysaccharide by inhalation. Inhal Toxicol 2006; 18:555–568.
- Wittmann I, SchÖnefeld M, Aichele D, Groer G, Gessner A, Schnare M. Murine bactericidal/permeability-increasing protein inhibits the endotoxic activity of lipopolysaccharide and gram-negative bacteria. J Immunol 2008; 180:7546–7552.
- Weiss J. Bactericidal/permeability-increasing protein (BPI): structure, function and regulation in host defense against Gram-negative bacteria. Biochem Soc Trans 2003; 31:785
- Chien JW, Zhao LP, Hansen JA, Fan WH, Parimon T, Clark JG. Genetic variation in bactericidal/permeability- increasing protein influences the risk of developing rapid airflow decline after hematopoietic cell transplantation. Blood 2006; 107:2200–2207.
- Levy O. A neutrophil-derived anti-infective molecule: Bactericidal/permeability-ßßincreasing protein. Antimicrob Agents Chemother 2000; 44:2925–2931.
- Ruiz A.G, Casafont F, Teran A, De La Pena J, Estebanez A, Romero FP. Increased bactericidal/permeability increasing protein in patients with cirrhosis. Liver Int 2010; 30:94–101.
- Groenewegen KH, Dentener MA, Wouters EFM. Longitudinal follow-up of systemic inflammation after acute exacerbatio of COPD. Respir Med 2007; 101:2409–2415.
- Hubacek JA, Stuber F, Frohlich D, Book M, Wetegrove S, Ritter M, Rothe G, Schmitz G. Gene variants of the bactericidal/permeability-increasing protein and lipopolysaccharide binding protein in sepsis patients. Crit Care Med 2001; 29:557–566.
- Hubacek JA, Aslanidis C, Buchler C, Schmitz G. PstI-polymorphismin the human bactericidal permeability increasing protein (BPI) gene. Clin Genet: 1979; 52:249.
- Maris NA, de Vos A.F, Dessing MC, Spek CA, Lutter R, Jansen HM, van der Zee JS, Bresser P, van der Poll T. (2005) Antiinflammatory effects of salmeterol after inhalation of lipopolysaccharide by healthy volunteers. Am J Respir Crit Care Med 2005; 172:878–884.
- Krawczak M, Reiss J, Cooper DN. The mutational spectum of single base-pair substitutions in mENA splice junctions of human genes: causes and consequences. Hum Genet 1992; 90:41–54.
- Cartegni L, Chew SL, Krainer AR. Listening to silence and understanding nonsense: Exonic mutations that affect splicing. Natrue Rev Genet 2002; 3:285–298.
- Klein W, Tromm A, Folwaczny C, Hagedorn M, Duerig N, Epplen J, Schmeiegel W, Griga T. A polymorphism of the bacterial/permeability protein gene is associated with Crohn's disease. J Clin Gastroenterol 2005; 39:282–283.
- Yang P, Tremanine WJ, Meyer RL, Prakash UB. Alpha 1-antitrypsin deficiency and inflammatory bowel diseases. Mayo Clin Proc 2000; 75: 450–455.
- Ekbom A, Brandt L, Granath F, Lofdahl C-G, Egesten A. Increased risk of both ulcerative colitis and Crohn's disease in a population suffering from COPD. Lung 2008; 186: 167–172.
- Sinden NJ, Stockley RA. Systemic inflammation and comorbidity in COPD: a result of overspill of inflammatory mediators from the lungs? Review of the evidence. Thorax 2010; 65:930–936.
- Eagan TML, Ueland T, Wagner PD, Hardie JA, Mollnes TE, Damas JK, Aukrust P, Bakke PS. Systemic inflammatory markers in COPD: results from the Bergen COPD cohort study. Eur Respir J 2010; 35:540–548.
- Walter RE, Wilk JB, Larson MG, Vasan RS, Keaney JF, Lipinska I, O'Connor GT, Benjamin EJ. Systemic inflammation and COPD. The Framingham heart study. Chest 2008; 133: 19–25.
- Levin M, Quint PA, Goldstein B, Barton P, Bradley JS, Shemie SD, Yeh T, Kim SS, Cafaro DP, Scannon PJ, Giroir BP. Recombinant bactericial/permeability- increasing protein (rBPI21) as adjunctive treatment for children with severe meningococcal sepsis: a randomized trial. Lancet 2000; 356:961–967.