The Use of Pulse Oximetry to Determine Hypoxemia in Acute Exacerbations of COPD

Abstract

Background: There is little evidence that the guideline-recommended oxygen saturation of 92% is the best cut-off point for detecting hypoxemia in COPD exacerbations. Objective: To detect and validate pulse oximetry oxygen saturation cut-off values likely to detect hypoxemia in patients with aeCOPD, to explore the correlation between oxygen saturation measured by pulse oximetry and hypoxemia or hypercapnic respiratory failure. Methodology: Cross-sectional study nested in the IRYSS-COPD study with 2,181 episodes of aeCOPD recruited between 2008 and 2010 in 16 hospitals belonging to the Spanish Public Health System. Data collected include determination of oxygen saturation by pulse oximetry upon arrival in the emergency department (ED), first arterial blood gasometry values, sociodemographic information, background medical history and clinical variables upon ED arrival. Logistic regression models were performed using as the dependent variables hypoxemia (PaO2 < 60 mmHg) and hypercapnic respiratory failure (PaO2 < 60 mmHg and PaCO2 > 45). Optimal cut-off points were calculated. Results: The correlation coefficient between oxygen saturation and pO2 measured by arterial blood gasometry was 0.89. The area under the curve (AUC) for the hypoxemia model was 0.97 (0.96–0.98) and the optimal cut-off point for hypoxemia was an oxygen saturation of 90%. The AUC for hypercapnic respiratory failure was 0.90 (0.87–0.92) and the optimal cut-off point was an oxygen saturation of 88%. Conclusions: Our results support current recommendations for ordering blood gasometry based on pulse oximetry oxygen saturation cut-offs for hypoxemia. We also provide easy to use formulae to calculate pO2 from oxygen saturation measured by pulse oximetry.

Keywords:

• diagnosis
• emergency department
• respiratory failure
• validation

Background

The Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines recommend performing arterial blood gasometry when clinical signs of respiratory failure or right heart failure are observed in patients with stable severe COPD. Arterial blood gasometry is also recommended for tracking or managing oxygen therapy in patients experiencing acute exacerbations of COPD, and when respiratory failure is suspected in order to initiate mechanical ventilation (1,2).

Until the results of blood gasometry are available, pulse oximetry is the screening tool used by physicians to make decisions about acute exacerbations of COPD (aeCOPD). The cut-off values for oxygen saturation as ­measured by pulse oximetry to detect hypoxemia ranges in different studies from 88% to 92%, depending on the amount of carboxyhemoglobin measured in peripheral blood samples (3). Relatively little is known about the range of oxygen saturation measured by pulse oximetry more related to hypoxemia in patients with aeCOPD. Yet, recommendations for oxygen treatment have been made based on low levels of evidence (4) or articles with small sample sizes (1,5). Although guidelines tend not to replace arterial blood gasometry for detecting hypoxemia, pulse oximetry is recommended as a useful screening tool : guidelines for the management and treatment of acute exacerbations of COPD have set recommendations for oxygen therapy based on values of oxygen saturation measured by arterial blood gasometry.

Pulse oximetry has been, therefore, used in emergency departments and primary care offices to diagnose hypoxemia as well as to manage and monitor patients with chronic obstructive pulmonary disease (COPD) and other respiratory conditions when stable or when experiencing acute exacerbations since this tool assesses oxygenation (1,4,6). On the other hand, although capnography has been developed to assess ventilation in emergency settings, this tool is generally not used in pre-hospital settings or primary care, and is not even available in all emergency departments.

By this reason, decisions about the initial treatment of hypoxemia before the results of arterial blood gasometry are available are based almost exclusively on pulse oximetry even though one of its key limitations is its inability to assess ventilation in the presence of oxygen therapy, which is used by many people with chronic respiratory conditions (6,7). It is known that the use of oxygen therapy in situations of ventilatory failure in acute excaerbations of COPD (aeCOPD) many times contributes to admission to the intensive care unit and increased length of stay (8).

In this context, the aims of this study were to 1) detect or validate cut-off values of oxygen saturation determined by pulse oximetry that are more likely to detect hypoxemia, 2) determine if oxygen saturation measured by pulse oximetry is correlated with hypoxemia or respiratory failure, and 3) identify possible confounders in this relationship.

Methods

We conducted a cross-sectional cohort study nested in the Investigacion en Resultados y Servicios Sanitarios-Chronic Obstructive Pulmonary Disease (IRYSS-COPD) appropriateness study, which has been described elsewhere (9). Patients for this cohort study were recruited from 16 hospitals belonging to the Spanish National Health Service. Patients with symptoms of COPD exacerbation who attended the EDs of these hospitals between June 2008 and September 2010 were invited to participate in the study. The Institutional Review Board of each hospital approved the study. All patients or caregivers had to provide informed consent to participate in the study.

Patients were candidates for the study if they presented to the ED of any of the participating hospitals with symptoms consistent of an exacerbation of COPD. Exacerbation was defined as an event in the natural course of the disease characterized by a change in the patient's baseline dyspnea, cough, and/or sputum that was beyond normal day-to-day variations, was acute in onset, and may have warranted a change in regular medication in a patient with underlying COPD (1).

Patients were considered to have been previously diagnosed with COPD if they had a forced expiratory volume in 1 second (FEV₁)/forced vital capacity (FVC) quotient <70%. Patients not previously diagnosed with COPD but in whom the disease was suspected were also eligible for inclusion in the study. This included smokers or former smokers of more than 15 packs per year with dyspnea, cough, or expectoration for more than 3 months per year, and who were experiencing symptoms resembling aeCOPD. The diagnosis had to be confirmed by spirometry within 60 days after the index episode at a time when the patient was stable, i.e., the absence of any increase in symptoms or changes in background therapy. If a diagnosis of COPD was not confirmed, the patient was excluded from the study.

Patients were excluded from the study if they had COPD complicated by a comorbidity such as pneumonia, pneumothorax, or pulmonary embolism; lung cancer; or left cardiac insufficiency. Other exclusion criteria included a diagnosis of asthma, extensive bronchiectasis, sequelae of tuberculosis, pleural thickening, or restrictive diseases. Patients who did not wish to participate were also excluded.

We collected information about patients upon their arrival in the ED (ED arrival). This information included the severity of the exacerbation (gauged by level of consciousness, respiratory rate, heart rate, and blood gasometry)(10); baseline severity of COPD (measured by the FEV₁)(1); the presence of comorbidities that affect the exacerbation of COPD, such as diabetes and cardiovascular disease; the number of hospital admissions for COPD exacerbations during the previous 12 months; the patient's ability to carry out necessary treatments at home; response to previous treatments; previous use of home oxygen therapy; and demographic variables such as age and sex. We also collected information about comorbidities included in the Charlson Co-morbidity Index. “Low perfusion” was created as a variable composed by heart rate, blood pressure and presence of arrhythmia: if systolic blood pressure lower than 90 mm Hg and heart rate lower or equal to 60 or greater or equal than 100 and/or presence of arrhythmia, then low perfusion is assumed; otherwise, low perfusion is not assumed.

We defined hypoxemia as pO2 < 60 mm Hg (7.99 kPa). Type II hypercapnic respiratory failure was defined as pO2 < 60 mm Hg (7.99 kPa) and pCO2 > 50 mm Hg (6.66 kPa), while type I respiratory failure was defined as pO₂ < 60 mm Hg (7.99 kPa) and pCO2 ≤ 50 mm Hg (6.66 kPa). A pCO2 > 50 mm Hg (6.66 kPa) was defined as hypercapnia.

Statistical analysis

Only aeCOPD episodes in which the oxygen level was measured with both pulse oximetry and arterial blood gasometry were included in the analysis. We excluded aeCOPD episodes in which oxygen saturation upon ED arrival were >92%, for which gasometry is not always required (1). Patients who presented with highly discordant values, such as arterial pO2 < 60 mm Hg and oxygen saturation measured by pulse oximetry > 90% or those with pO2 > 100 mmHg were also excluded.

Descriptive statistics include frequencies and percentages for categorical variables and mean and standard deviation (SD) for continuous variables. The sample was randomly divided into two equal-sized subsamples: a derivation sample, composed of 1090 episodes of COPD exacerbations, and a validation sample, composed of 1091 episodes.

The derivation sample was used to develop a model that describes the relationship between oxygen saturation measured by pulse oximetry and pO2 measured by blood gasometry. The Spearman correlation coefficient between both measures was calculated. The relationship between the two was shown in a scatter plot. As the relationship was not linear, the x axis (pulse oximetry) was divided into sections of the same length and the functional relationship was adjusted using locally weighted polynomial regression.

We compared the adjacent slopes of the polynomial pieces, looking for the point where the maximum slope change occurred. Bootstrap resampling techniques were used to test statistical significance of the slope change at different cut-off points to validate the selected point at which the maximum slope change occurs. Two different regression lines were then calculated, using that point as the cut-off, by performing a multilevel analysis adjusted by hospitals. Finally, both regression models were combined into a prediction rule for pO₂ detected by blood gasometry using as a predictor O₂ saturation by pulse oximetry. The predictive rule from the derivation cohort was then applied to the validation sample and the value of R² was calculated as a measure of variability explained by the models.

Logistic regression models were created to explore the probability of detecting hypoxemia and hypercapnic respiratory failure. In addition to oxygen saturation measured by pulse oximetry, other independent variables considered were baseline severity of COPD, age, gender, presence of arrhythmia, blood pressure, smoking habit, hospital admissions for aeCOPD in the previous year, and heart rate and respiratory rate upon ED arrival.

The predictive validity of each model was determined by calculating the area under the receiver operating characteristic (ROC) curve (AUC). Characteristics of the test were evaluated by sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), as well as accuracy. For each model, the point that maximized the sum of sensitivity and specificity was considered the optimal cut-off point.

These models were then applied to the validation sample to compare predictive accuracy by means of comparisons of AUCs, and the cut-off points were validated by analysing sensitivity and specificity (11).

All effects were considered significant at P < 0.05 unless otherwise stated. All statistical analyses were performed using SAS for Windows statistical software, version 9.2 (SAS Institute, Inc., Carey, NC) and R^© software version 2.13.0.

Results

We recruited 2877 episodes of COPD exacerbations. We excluded 433 by the reason of missing data in pO2 or in Sat O2 and 181 episodes because presenting discordant values between pulse oxymerey and gasometry, that is, SatO2 > 90 and, at the same time, pO2 < 60. We also excluded 82 patients who presented pO2 > 100. We have excluded by all these reasons 696 episodes of COPD exacerbations.

The analysis includes 2,181 episodes of COPD exacerbation. Of these, 604 episodes (27.71%) presented with type 1 respiratory failure, 472 (21.65%) with type 2, and 693 (31.79) with hypercapnia. The mean oxygen saturation measured by pulse oximetry was 87.28% (±9.33)and pO₂ measured by arterial blood gasometry was 59.3 mm Hg (±14.11). Nearly half of the patients (1077 [49.38%]) had significant systemic hypoxemia. Complete characteristics of the derivation and validation cohorts are presented in Table 1.

Table 1.  Descriptive analysis

	Derivation		Validation
	N	n (%)	N	n (%)	p-value
Age*	1090	72.82(9.61)	1091	73.14(9.35)	0.61
Sex -Male	1090	1002(91.93)	1091	978(89.64)	0.0652
Basal FEV	931		912		0.6796
≥50		308(33.08)		310(33.99)
<50		623(66.92)		602(66.01)
Charlson Index*	1090	2.29(1.62)	1091	2.22(1.51)	0.7813
Previous home oxygen therapy or nor invasive mechanical ventilationat home	1090	410(37.61)	1091	425(38.96)	0.5196
Nº previous admissions due to COPD-3 or more	1079	138(12.79)	1072	131(12.22)	0.6897
Edema – Yes	1037	204(19.67)	1038	221(21.29)	0.3609
Dyspnoea at rest at arrival to ED - Yes	1054	734(69.64)	1034	717(69.34)	0.8828
Cardiopathy	1083	223(20.59)	1082	238(22)	0.4245
Arrhytmia	721	119(16.5)	684	140(20.47)	0.0555
Diabetes	1083	242(22.35)	1084	231(21.31)	0.5596
Continuous positive airway pressure	1072	88 (8.21)	1073	84 (7.83)	0.7456
Severity of COPD at baseline	931		912		0.7598
Mild		32(3.44)		28(3.07)
Moderate		261(28.03)		268(29.39)
Severe		638(68.53)		616(67.54)
Severity of COPD exacerbation	572		605		0.3767
Mild		205(35.84)		231(38.18)
Moderate		221(38.64)		210(34.71)
Severe		146(25.52)		164(27.11)
Low perfusion	604	15(2.48)	543	16(2.95)	0.6291
pO2*	1090	59.12(13.85)	1091	59.49(14.38)	0.4960
pCO2*	1089	47.96(14.06)	1091	47.46(13.39)	0.2020
pH*	1089	7.4(0.06)	1090	7.4(0.06)	0.7147
O2 Saturation*	1090	87.36(8.97)	1091	87.2(9.67)	0.6738
Hypoxemia	1090	543(49.82)	1091	534(48.95)	0.6843
Respiratory failure Type I	1089	303(27.82)	1091	301(27.59)	0.9027
Respiratory failure Type II	1089	239(21.95)	1091	233(21.36)	0.738
Hipercapnia	1089	355(32.6)	1091	338(30.98)	0.4173

*Represented as mean (std).

The correlation coefficient between oxygen saturation measured by pulse oximetry and pO₂ determined from arterial blood gasometry was 0.89. The relationship the two measured is shown in Figure 1. The greater change in slope was detected where O₂ saturation is >90%. The resulted equations were:

if SatO₂ ≥ 90 then pO₂ = (SatO₂ × 0.7712) − 13.41

if SatO₂ > 90 then pO₂ = (SatO₂ × 3.3235) − 241.65

Figure 1. Relationship between oxygen saturation by pulse oximetry and calculated from blood gas analysis.

This model explained 74% of the variability in pO₂.

The ROC curve for determining the oxygen saturation screening cut-off for systemic hypoxemia by pulse oximetry is shown in Figure 2. The area under the curve for the model is 0.97 (95% confidence interval [95% CI], 0.96–0.98). The optimal cut-off suggested by this curve is at a pulse oximetry oxygen saturation of 90%. That corresponds with a sensitivity of 100% and a specificity of 84%. Candidate variables included in the model that were significantly associated with oxygen saturation were baseline severity of COPD and dyspnea at rest upon ED arrival. Age, sex, number of hospital admissions for a COPD exacerbation in the previous year, and respiratory rate upon ED arrival were not significantly associated with oxygen saturation. The performance of different screening cut-offs is shown in Table 2.

Figure 2. ROC curve for ability of oxygen saturation by pulse oximetry predicting systemic hypoxemia.

Table 2.  Performance of different levels of screening cut-off for the detection of hypoxia by pulse oximetry in derivation sample

Oxygen saturation by pulse oximetry (%)	Accuracy	Sensitivity for detection of hypoxia	Specificity	Positive predictive value	Negative predictive value
87	0.82	0.69	0.96	0.95	0.76
88	0.86	0.78	0.94	0.93	0.81
89	0.90	0.89	0.91	0.90	0.89
90	0.92	1	0.84	0.86	1
91	0.89	1	0.77	0.81	1
92	0.83	1	0.65	0.74	1

The ROC curve for determining the oxygen saturation screening cut-off for hypercapnic respiratory failure by pulse oximetry is shown in Figure 3. The area under the curve is 0.90 (95% CI, 0.87–0.92). The optimal cut-off suggested by this curve is at pulse oximetry oxygen saturation of ≤88%. That corresponds with a sensitivity of 91% and specificity of 72%. The performance of different screening cut-offs is shown in Table 3.

Figure 3. ROC curve for the performance of oxygen saturation by pulse oximetry for predicting hypercapnic respiratory failure.

Table 3.  Performance of different levels of screening cut-off for the detection of hypercapnic respiratory failure by pulse oximetry

Oxygen saturation by pulse oximetry (%)	Accuracy	Sensitivity for detection of hypoxia	Specificity	Positive predictive value	Negative predictive value
87	0.79	0.85	0.78	0.51	0.97
88	0.72	0.91	0.72	0.48	0.97
89	0.71	0.95	0.64	0.43	0.98
90	0.64	1	0.54	0.38	1
91	0.61	1	0.50	0.36	1
92	0.55	1	0.42	0.33	1

Discussion

In our sample of 2,181 episodes of COPD exacerbations evaluated in 16 different hospitals, the optimal pulse oximetry cut-off point for detecting hypoxemia was an oxygen saturation of 90%, while the optimal cut-off to detect hypercapnic respiratory failure was 88%. These findings support in part GOLD recommendations to assess arterial blood gases when oxygen saturation measured by pulse oximetry is below 92%, and provide information to clinicians about the probability of hypercapnic respiratory failure when oxygen saturation is below 88%.

The GOLD guidelines are widely used in Spain and other countries. However, it must be noted that the GOLD recommendation about the use of gasometry in aeCOPD is based on just one article, which was published in Respiratory Medicine in 2001 (5,7), when the first version of the guideline was published. That article reports results from just 64 sample pairs of oxygen saturation determined by pulse oximetry and pO2 measured by blood gasometry. In 2006, Guryay et al. examined the capability of pulse oximetry to detect hypoxemia and also measured its ability to detect hypercapnia in a sample of 76 patients with exacerbations of COPD (12). They proposed a cut-off point of 88.5% as an indication for hypoxemia, which is clearly lower than cut-off of 90% that we propose. The small sample size of Guryay et al. limits their recommendation. In addition, the sensitivity of their test to detect hypercapnia (73%) was lower than sensitivity we observed (91%).

To our knowledge, no other publications to date have measured the accuracy of pulse oximetry to predict hypoxemia and/or respiratory failure, even though pulse oximetry is widely used to assess the severity of COPD exacerbations and managing diagnosis, oxygen therapy, and ventilatory support (13–16).

Since our study is observational in nature, its results cannot definitively determine that an oxygen saturation below 92% is warranted to order blood gasometry (1). Nevertheless, our findings raise two important points for clinicians:

1. The relationship between oxygen saturation measured by pulse oxymetry and pO2 measured by arterial blood gasometry is not linear. This could be important to take into account for patients experiencing acute exacerbations of COPD. As oxygen saturation increases, the probability of increasing pO2 increases exponentially .The British Thoracic Society guideline for the use of oxygen therapy in adults at risk of respiratory failure establishes a target oxygen saturation range of 88% to 92% (4). Our results provide easily used formulas to estimate pO₂ based on oxygen saturation measured by pulse oximetry.

2. We encountered different cut-off points for detecting hypoxemia and hypercapnia in our sample. This finding, if confirmed in other studies, could change the decision to delay oxygen therapy until blood gases have been evaluated and to offer individualized doses of oxygen therapy. It is well known that intensive oxygen therapy in patients experiencing COPD exacerbations who are at risk of hypercapnic respiratory can be hazardous, with the use of high-flow oxygen contributing to admission to the intensive care unit (8) and increased length of stay. For that reason we believe that a change in the hypercapnia detection cut-off point to an oxygen saturation determined by pulse oximetry of 88% could delay or decrease the use of oxygen therapy in these patients. As Witting et al. related in 2001 (6), pulse oximetry could detect moderate hypercapnia among patients breathing normal air.

Authors have given different cut off points in order to help decision over these patients. We calculated optimal cut off point for detection of hypercapnic respiratory failure as the point which maximized the sum of sensitivity and specificity. The decision maker, however, can choose the cut-off closest to her/his interest in each specific patient or situation.

The large number of patients and episodes of COPD exacerbations we recruited for this study is an important strength. Sample sizes of similar prior studies were under 100; our sample size was 20-fold higher. In addition, this was a real-life study in which no efforts were made to influence how clinicians treated their patients. This makes the results more generalizable. We also provide easy-to-use formulas for calculating pO2 at the bedside based on oxygen saturation determined by pulse oximetry, which could serve to provide individualized doses of oxygen therapy to patients experiencing COPD exacerbations.

One limitation of our study is the absence of variables related to the probability of respiratory failure, such as acute congestive heart failure, chest wall deformities, carboxyhemoglobin intoxications, etc. Thus we were not able to measure their influence on oxygen saturation. Nevertheless, an intermediate variable of low perfusion (a key reason why pulse oximetry may fail to detect hypoxemia) has been explored in multivariate analysis. We defined this variable depending on blood pressure, cardiac rate and presence of arrhythmia. We also explored whether variables such as baseline severity of COPD, age, gender, and smoking habit significantly affected oxygen saturation cut-offs when stratified analysis was performed. Lee et al. recommended caution when using pulse oximetry as screening tool to detect hypoxemia in patients likely to have carboxyhemoglobin ≥2% (3). Other authors have found no influence of anemia, acidosis, hyperlactatemia, or vasoactive drugs on results of pulse oximetry (17,18). Our findings indicate that pulse oximetry provides a useful way to detect hypoxemia, although future studies must evaluate how well it works in different subgroups of patients and in different situations.

Conclusions

Our results confirm pulse oximetry oxygen saturation cut-off points recommended by the GOLD guidelines for ordering arterial blood gas measurements in patients experiencing acute exacerbations of COPD. Because our study was observational, it should not be used to direct clinical decision making. Yet, the results should serve as the basis for developing future studies in controlled conditions to validate our cut-offs and compare the results obtained by pulse oximetry and capnography. Further work will be required to confirm the oxygen saturation cut-off we recommend for detecting hypercapnia in patients experiencing acute exacerbations of COPD.

Declaration of Interests

This work was supported in part by grants from the Fondo de Investigación Sanitaria (PI 061010, PI061017, PI06714, PI060326, PI060664); Department of Health of the Basque Country; and the thematic ­networks- Red IRYSS (Investigacion en Resultados y ­Servicios Sanitarios)- of the Instituto de Salud Carlos III (G03/220), and the European Regional Development Fund (ERDF).

No possible conflicts of interest (e.g., funding sources for consultancies or studies of products) exist in this study. The authors alone are responsible for the content and writing of the paper.

Acknowledgments

We acknowledge editorial assistance of Patrick Skerret. The IRYSS- COPD group included the following co-investigators: Dr. Jesús Martínez-Tapias (Hospital Virgen de las Nieves, Granada); Alba Ruiz (Hospital de Motril, Granada); Dr. Eduardo Briones (Unidad de Epidemiología. Distrito Sanitario Sevilla); Dra. Silvia Vidal (Unidad de Investigación, Hospital Costa del Sol, Marbella); Dr. Emilio Perea-Milla, Francisco Rivas (Servicio de Epidemiología, Hospital Costa del Sol, Málaga – REDISSEC); Dr. Maximino Redondo (Servicio de Laboratorio, Hospital Costa del Sol, Málaga-REDISSEC); Javier Rodríguez Ruiz (Responsable de Enfermería del Área de Urgencias, Hospital Costa del Sol, Málaga); Dra. Marisa Baré (Epidemiología y Evaluación, Corporació Sanitaria Parc Taulí-CSPT, Sabadell REDISSEC), Dr. Manel Lujan, Dra. Concepción Montón (Servicio de Neumología, CSPT/REDISSEC); Dra. Amalia Moreno, Dra. Josune Ormaza, Dr. Javier Pomares (Servicio de Neumología, CSPT); Dr. Juli Font (Medicina, Servicio de Urgencias; CSPT), Dra. Cristina Estirado, Dr. Joaquín Gea (­Servicio de Neumología, Hospital del Mar/CIBERES, Barcelona); Dra. Elena Andradas (subdirectora de Promoción de la Salud y Epidemiología del Ministerio de Sanidad, ­Servicios Sociales e Igualdad), Dr. Juan Antonio Blasco (Unidad de Evaluación de ­Tecnologías Sanitarias, ­Agencia Laín Entralgo, Madrid), Dra. Nerea Fernández de ­Larrea (Subdirección General de Tecnología e Innovación Sanitarias. Consejería de Sanidad de la Comunidad de Madrid/REDISSEC); Rosa Girón (Hospital de La ­Princesa, Madrid), María del Puerto Cano Aguirre (Hospital de Torrejón, Madrid); Dr. Jose Luis Lobo (­Servicio de Neumología, Hospital Txagorritxu, Araba); Dra. Esther Pulido (Servicio de Urgencias, Hospital Galdakao-­Usansolo, Bizkaia); Dr. Mikel Sánchez (­Servicio de ­Urgencias, Hospital Galdakao-Usansolo); Dr. Luis Alberto Ruiz (Servicio de Respiratorio, Hospital de Cruces, Bizkaia); Dra. Ane Miren Gastaminza (Hospital San Eloy, Bizkaia); Dra. Eva Tabernero (Servicio de ­Neumología, Hospital de Santa Marina), Carmen Haro (Servicio de Urgencias, Hospital de Santa Marina, Bizkaia); Dr. Ramon Agüero (­Servicio de Neumología, Hospital Marques de Valdecilla, Santander); Dr. Gabriel Gutiérrez (Servicio de Urgencias, Hospital Cruces, ­Bizkaia); Dra. Belén Elizalde (Dirección Territorial de Gipuzkoa); Dr. Felipe Aizpuru (Unidad de Investigación, Hospital Txagorritxu, Álava/REDISSEC); Dra. Inmaculada Arostegui, Irantzu Barrio (Departamento de Matemática Aplicada, Estadística e Investigación Operativa, UPV/EHU- REDISSEC; Amaia Bilbao (Hospital Universitario Basurto/REDISSEC); Dr. Cristóbal Esteban (Servicio de Neumología, Hospital Galdakao-Usansolo, Bizkaia/REDISSEC); Dra. Nerea González, Susana ­Garcia, Iratxe Lafuente, Urko Aguirre; Miren Orive, Ane Anton, Dr. Jose M. Quintana (Unidad de Investigación, Hospital Galdakao-Usansolo, Bizkaia/REDISSEC).