Peak expiratory flow rate as predictor of inpatient death in patients with chronic obstructive pulmonary disease.

Objectives: Few studies analyze hospital deaths and related factors in patients with acute exacerbation of chronic obstructive pulmonary disease who require hospitalization.

Methods: A cross-sectional study was done with 284 patients who had been admitted consecutively to the Short Stay Medical Unit at the Juan Canalejo Hospital in A Coruna.

Results: Eleven patients (3.9%) died. The independent variables for predicting death were the peak expiratory flow (OR, 0.96; 95% CI, 0.94 to 0.98), long-term oxygen therapy (OR, 12.46; 95% CI, 2.1 to 72.4), and body mass index (OR, 0.73; 95% CI, 0.59 to 0.90). A peak expiratory flow < 150 L/min showed the best specificity and positive predictive value with maximum sensitivity for predicting death. The results of the arterial blood gasses and the functional tests did not predict hospital death.

Conclusions: Peak expiratory flow was the most important predictive value for determining the risk of death in patients who required hospitalization for acute exacerbation of chronic obstructive pulmonary disease. Additional studies are required to validate these findings.

Key Words: chronic obstructive pulmonary disease, inpatient death, logistic regression, peak expiratory flow rate, receiver operating characteristic curves 

**********

Chronic obstructive pulmonary disease (COPD) affects 15 million Americans and is responsible for more than 160,000 deaths per year, ranking fourth as the cause of death in the United States. (1,2) In Spain, the prevalence of COPD in people between 40 and 69 years of age is 9%, and it is ranked fifth as a cause of death among males and eighth among females. (3) The disease incurs exorbitant economic costs. The total annual cost of COPD care in the United States is $6.6 billion. (4)

Sixty percent of these patients die within 10 years of being diagnosed. (5) Although not all studies agree, various factors predicting death have been identified, such as age, sex, a history of smoking, body mass index, and functional, electrocardiographic, and hemodynamic parameters. (5-12)

Few studies analyze the factors predicting death in patients with an exacerbation of their COPD for which they require hospitalization. (13,14) The present study was designed to determine the prognostic value of peak expiratory flow (PEF) in these patients.

Materials and Methods

Patients and study design

A prospective study was conducted on 284 patients who had been admitted consecutively to the Short Stay Medical Unit (SSMU) at the Juan Canalejo Hospital in A Coruna (Spain) from February 2000 to March 2001. The SSMU is a 41-bed unit within a 1,100-bed university teaching hospital, and it is staffed by hospital-based internists. Over the 1-year study period, 3,000 patients were admitted to the SSMU, and the average length of stay was 4 days. The three most frequent principal discharge diagnoses were heart failure, COPD exacerbation, and coronary artery disease. COPD diagnosis was made according to the Spanish Society of Pneumology and Thoracic Surgery (15) and the GOLD standards. (16) For this, it was considered essential to demonstrate chronic airflow limitation defined by a postbronchodilator FE[V.sub.1] less than 80% of the predicted value in combination with an FE[V.sub.1]/forced vital capacity (FVC) of less than 70%. (16)

Our criteria for hospitalization were similar to those of the expert consensus of the American Thoracic Society. (17) Hospitalizations in preparation for planned invasive surgical or diagnostic procedures were excluded from analysis. Patients were admitted from the emergency room to the SSMU without initial bias.

During the first 24 hours after admission, we recorded demographic and socioeconomic data, medical history, respiratory risk factors, and current clinical status, based on a questionnaire. In most cases, this questionnaire was answered by the patients themselves and by a relative when this was not possible (7.3%).

In patients who smoked, tobacco use was evaluated on an accumulative basis, expressed in pack-years. (18) The scale of the British Medical Research Council, as modified by the American Thoracic Society, (19) was used to determine the degree of dyspnea exhibited by the patients, and the Pratt criteria (20) were used in the radiologic diagnosis of emphysema. The diagnosis of cor pulmonale was made according to history, physical examination, and roentgenographic and electrocardiographic findings. (11,21) Echocardiography performed by various operators produced good-quality echocardiograms in only 40% of the patients. Accordingly, we excluded echocardiography as the sole criteria for diagnosing cor pulmonale. The PEF rate was measured at admission to the SSMU after initiating bronchodilator therapy in the emergency department, always by one of the authors, with results expressed in liters per minute. For this test, done at the patient's bedside by means of a portable Vitalograph meter (Vitalograph Ltd, Buckingham, UK), the best value of three consecutive attempts was taken.

We recorded the arterial gas analysis in the emergency room ([FIO.sub.2], 21%). The cause of exacerbation was classified according to European Respiratory Society standards. (22) The values of the functional tests and arterial blood gasses (inspired oxygen concentration, 21%) in a basal situation were obtained. Tests were considered valid if they had been done within the 4-month period before the admission, provided that the patient had remained in a stable condition for at least 2 weeks before the tests were performed. FE[V.sub.1] and FVC were measured with the use of a spirometer (Erich Jaeger GmbH & Co, KG, Wuerburg, Germany). The technique used to perform these tests complied with the international standard. (23) The spirometric reference values were those corresponding to a Mediterranean population. (24) Forty-three patients who had not been tested before admission were seen by appointment at the ambulatory clinic at least 1 month after they were discharged. If their situation had not stabilized, these studies were postponed. All studies were successfully completed within 3 months of discharge.

Statistical analyses

A descriptive study was conducted of all the variables included in the study with the corresponding 95% confidence interval (CI), and the results of patients who had died during admission were compared with those who survived.

The different averages were compared by means of the Student t test or the Mann-Whitney U test, as appropriate. The normalcy of the quantitative variables had been previously verified by using the Kolmogorov-Smirnov test. The association between qualitative variables was found by means of the [chi square] test, and the odds ratio (OR) was calculated with a 95% CI.

Calculations were made for sensitivity, specificity, and predictive values of PEF and various spirometric values; these parameters were calculated with their 95% CI, and the corresponding receiver operating characteristic (ROC) curves were drawn.

The prognostic factors were determined through a log regression analysis with the forward method in terms of the likelihood ratio. The dependent variable was death during hospitalization and the covariables were age, sex, a history of oncologic pathology or heart failure, and variables having a significant association evidenced in the univariate analysis. Qualitative variables with more than two response possibilities were categorized.

The significance level established in all the analyses was p less than 0.05, with all the tests being bilateral. Statistical processing was carried out with the use of the SPSS statistics pack version 6 for Windows (SPSS Inc, Chicago, IL).

Results

The average age of the patients was 71.3 [+ or -] 9.4 years (range, 44 to 97 years), with 75% of the group being older than 65 and 89.8% being male. Table 1 presents the most outstanding general characteristics of our cohort; Figure 1 shows the causes of exacerbation of COPD. Eleven patients died, which represents a hospital mortality rate of 3.9%. The causes of deaths were respiratory failure (4), pulmonary embolism (3), arrhythmia (1), pancreatic cancer (1), ischemic stroke (1), and mesenteric vascular occlusion (1). The average hospital stay was greater for patients who died (13.5 [+ or -] 11.5 versus 4.6 [+ or -] 5.1 days; P = 0.0108).

The univariate analysis showed significant differences in anthropometric and spirometric values, habitual therapy, degree of basal dyspnea, symptoms, signs, and value of PEF at admission (Table 2).

The validity and reliability of various functional parameters showed that the PEF was the one that obtained the greatest specificity and positive predictive value at the cutoff point for maximum sensitivity (Table 3).

The corresponding ROC curves were drawn. The greatest value in the area under the curve was obtained by the PEF (Fig. 2).

The multivariate analysis proved that the only variables that continued to be independent predictors of hospital death were long-term oxygen therapy, body mass index, and PEF, taking into account age, sex, weight, a history of heart failure or oncologic pathology, continuous oral corticosteroids, the degree of dyspnea exhibited by the patient, thoracic excursion values, FE[V.sub.1] and FE[V.sub.1]/FVC, number of days with increased symptoms, changes in expectoration, respiratory frequency, and the level of consciousness at admission (Table 4).

When a patient with COPD needs long-term oxygen therapy, he or she is at a greater risk of death during an exacerbation. Furthermore, this risk is inversely proportional to the body mass index and PEF values (Table 4 and Fig. 2).

Discussion

Our cohort of patients were mostly men, smokers, and from less favorable socioeconomic backgrounds, as commonly described in the literature. (15,17,22,25)

Hospital death caused by COPD exacerbation varies in accordance with the basal situation of the patients and the cause of the exacerbation. (13,14,26-28) Connors et al (14) reported a rate of 11% in a group of 1,016 patients, all with a PaC[O.sub.2] of 50 mm Hg or greater at the time of admission. Fuso et al (13) published a retrospective study that included 22% of the cases requiring mechanical ventilation, with a mortality rate of 14%. However, the mortality rate published by authors caring for patients hospitalized in conventional wards ranged between 4 and 6%, (26-28) which tallies with the 3.9% of our patients. In our cohort, most of the deaths were either directly related to COPD (respiratory failure, arrhythmia) or secondary to frequent complications developing in COPD patients (pulmonary embolism, ischemia).

If the factors predicting death were known, it would then be possible to select the cases of high risk to provide them more intensive therapy. The results of our study point to three variables: peak expiratory flow, long-term oxygen therapy, and body mass index.

In patients with COPD and chronic respiratory failure, oxygen therapy increases survival and modifies the natural evolution of the disease. (29,30) However, their survival rate is lower than that of the overall COPD patients. They live for an average of 3 years, and 50% died within a 5-year period (6,12); 27.5% of our patients were on domiciliary oxygen therapy and multiplied the risk of death during admission by 12.4 (95% CI, 2.1 to 72.4). We consider that these data reflect the severity of the disease in patients requiring long-term oxygen therapy.

A substantial proportion of COPD patients are undernourished, which has been related to death, even independent of the value of FE[V.sub.1]. (7,8,14,31) The OR for the body mass index in our study was 0.73 (95% CI, 0.59 to 0.90), which implies that a low index is associated with death and that with each unit increase of this index, the risk of death decreases 1.37 times. Patients in a worse nutritional condition have changes in their respiratory muscles, ventilatory drive, and immunologic system, all of which favor a poor evolution in cases of exacerbation. (32)

In our study, PEF was a predictor of death, as its value decreased (OR, 0.96; 95% CI, 0.94 to 0.98). For each higher unit (in liters per minute) shown by PEF at admission, the risk of death decreased 1.37 times. None of the patients who died presented with a PEF greater than 150 L/min, and no other data regarding either spirometrics or arterial blood gasses, in a basal or exacerbated situation, were capable of predicting death.

The measurement of PEF using a portable instrument is advantageous in that it can be done at the patient's bedside, with simple instructions and at low cost. Self-measured PEFs are reasonably precise and accurate in patients with COPD. (33) Some authors consider that it provides information that is comparable overall to that of spirometry. (17) However, it has a greater variability and is less sensitive to small functional changes in the airways. There are authors who do not find a clear correlation with the symptoms. (22,34,35) Murata et al (36) reported that certain PEF rate parameters and dyspnea scores are independent predictors of respiratory status in patients with COPD.

Emerman and Cydulka (37) compared PEF and FE[V.sub.1] values carried out in the emergency department on patients with acute COPD. They found a good correlation both in absolute values (r = 0.84; P < 0.001) as well as in expected percentages (r = 0.81; P < 0.001). A recent report studied the value of the peak flow as predictor of overall deaths in nonacutely exacerbated patients with COPD and asthma. After controlling for age, smoking, sex, and body mass index, Hansen et al (38) found the best PEF to be at least equal to the best FE[V.sub.1] as a predictor of overall deaths in subjects with COPD. This may be due to PEF and FE[V.sub.1] reflecting different components of COPD. Furthermore, extrapulmonary components such as muscle mass and general vigor probably affect PEF to a greater extent than they affect FE[V.sub.1].

The area under the ROC curve is an overall measure of the accuracy of a diagnostic test. (39) In our study, the PEF correctly classified 86% of our patients with regard to death during hospitalization. A PEF below 150 may be used as advice to admission to a bed that provides for continuous monitoring of pulse oximetry, capnography, and ECG and has ready access to facilities for intensive care.

We consider that the predictive capacity of PEF as opposed to the other functional respiratory tests is probably related to the proximity in time between the PEF measurement and the event of dying. The functional tests were carried out before or after admission to the hospital and always in a stable condition, which, on the other hand, was essential to ensure that the patients fulfilled the COPD diagnostic criteria.

Limitations of the Study

The capacity of the models of multivariate analysis for predicting death is affected by the reduced sample size of the deaths (11 deaths in our study). Despite this, we have found variables with significant predictive value, although the accuracy of one of them, long-term oxygen therapy, is scant, as the CI goes from 2 to 72. A larger cohort study set in different medical wards will probably validate these clinically significant findings.

[FIGURE 2 OMITTED]

A low PEF is found to predict an elevated mortality rate, better than baseline FE[V.sub.1] in stable condition. PEF and FE[V.sub.1] in this study had not been measured at the same time because FE[V.sub.1] values measured during a stable period is condition sine cua non for diagnosis COPD. (15-17)

Another limitation might have been a potential selection bias. However, the patients admitted to the SSMU were not preselected; 1,078 patients were admitted to Juan Canalejo Hospital for acute exacerbation of COPD from February 2000 to March 2001. No significant differences were found regarding age, sex, and death of the patients admitted to our unit versus other hospital units (71.3 [+ or -] 9 versus 70.8 [+ or -] 11 years, 89.8% versus 88.7% males, and 3.9% versus 4.2% deaths, respectively). Considering the above data, we think that this is a possible but not probable selection bias.

Conclusion

The factors that predict death during hospitalization for exacerbated COPD are a reduced PEF value at the time of admission, a low body mass index, and the need for long-term oxygen therapy. The PEF value correctly classifies 86% of the patients selected at random with regard to death during admission, and a value of 150 L/min or less suggests a high-risk patient.

Table 1. Summary of main characteristics of the patients

                            Mean (SD)    n (%)       Range    95% CI

Demographic variables
  Age, yr                    71.3 (9.4)              44-97
  Sex
    Female                                29 (10.2)            7.0-14.1
    Male                                 255 (89.8)           85.8-92.9
  Weight, kg                 71.6 (13)               31-136
  Body mass index, kg/       27.2 (4.8)              14.5-47
    [m.sup.2]
  Smoker, current and                    242 (85.2)           80.7-88.9
    former
Socioeconomic data
  Illiterate                              51 (18.0)           13.8-22.7
  Primary studies                        219 (87.0)           66.6-88.1
  Secondary or university                 14 (5.0)             2.1-9.5
    studies
  Family income/yr < 9,000               191 (67.3)           61.3-72.5
    Eur.
  Family income/yr > 9,000                93 (32.7)           26.5-40.2
    Eur.
Medical history
  Hypertension                            88 (31.0)           25.8-36.5
  Diabetes mellitus                       50 (17.6)           13.5-22.6
  Ischemic heart disease                  37 (13.0)            9.5-17.3
  Heart failure                           27 (9.5)             6.5-13.3
  Oncologic pathology                     26 (9.2)             6.2-12.9
Basal situation
  Dyspnea (MRC scale)
    Grade 0                                9 (3.2)             1.5-5.7
    Grade 1                               53 (18.7)           14.4-23.5
    Grade 2                              122 (43.0)           37.2-48.7
    Grade 3                               49 (17.3)           13.1-21.9
    Grade 4                               51 (18.0)           13.8-22.7
  Cor pulmonale                          110 (38.7)           33.2-44.5
  Pa[O.sub.2], mm Hg         66 (12)                 34-99
  PaC[O.sub.2], mm Hg        43 (7)                  25-69
  FE[V.sub.1] %              43 (16)                 17-79
  FE[V.sub.1]/FVC            47 (11)                 21-70
Habitual therapy
  Long-term oxygen therapy                78 (27.5)           22.5-32.8
  Continuous oral                         59 (20.8)           16.3-25.7
    corticoids
Admission data
  Changes in expectoration               260 (91.5)           87.8-94.3
  Increased symptoms, d       3.3 (2.8)               1-30
  Heart rate                 94.9 (18.3)             36-200
  Respiratory rate           27.3 (6)                14-60
  Thoracic excursion, cm      2.1 (1.4)               0-8
  Pa[O.sub.2], mm Hg         54 (11)                 28-88
  PaC[O.sub.2], mm Hg        46 (12)                 21-98
  PEF, L/min                169 (71)                 20-400
Mean stay, d                  4.9 (5.7)               1-64
In-patient mortality                      11 (3.9)             2.0-6.6

95% CI, 95% confidence interval; PEF, peak expiratory flow; Eur., Euros.

Infection     83.5%
Pneumonia      5.6%
HF/Arrhythm.   8.5%
P.E.           0.4%
End stage      0.7%
Others*        1.4%

HF: heart failure. Arrhythm.: arrhythmia. P.E.: pulmonary embolism.
*: allergic reaction, vertigo, pulmonary tuberculosis and upper
gastrointestinal bleeding.

Fig. 1 Cause of exacerbation of chronic obstructive pulmonary disease
According to the European Respiratory Society (percentage of patients).

Note: Table made from bar graph.

Table 2. Univariate analysis of mortality (a)

                                          Living
                                 Mean (SD)     n (%)

Demographic variables
  Age, yr                         71.1 (9.4)
  Sex
    Male                                       247 (90.5)
    Female                                      26 (9.5)
  Weight, kg                      72.0 (13.8)
  Body mass index, kg/[m.sup.2]   27.3 (4.8)
Medical history
  Heart failure                                 25 (9.2)
  Oncologic pathology                           24 (8.8)
Basal situation
  Dyspnea
    Grade 0                                      9 (3.3)
    Grade 1                                     53 (19.4)
    Grade 2                                    122 (44.7)
    Grade 3                                     46 (16.8)
    Grade 4                                     43 (15.8)
  Cor pulmonale                                104 (38.1)
  Pa[O.sub.2], mm Hg              66.8 (12)
  PaC[O.sub.2], mm Hg             43.6 (7)
  FE[V.sub.1], %                  44.4 (16)
  FE[V.sub.1]/FVC                 47.9 (11)
Habitual therapy
  Long-term oxygen therapy                      70 (25.6)
  Continuous oral corticoids                    54 (19.8)
Admission data
  Changes in expectoration                     252 (92.3)
  Increased symptoms, d            3.1 (2.3)
  Respiratory rate                27.0 (5.5)
  Thoracic excursion, cm           2.2 (1.4)
  Level of consciousness
  Alert                                        254 (93.0)
  Confusion                                     17 (6.2)
  Stupor                                         2 (0.7)
  Coma                                           0
  Pa[O.sub.2], mm Hg*             54.3 (11)
  PaC[O.sub.2], mm Hg*            46.1 (12)
  PEF, L/min                     173.2 (69)

                                          Dead
                                 Mean (SD)     n (%)     P

Demographic variables
  Age, yr                        75.5 (8.8)              0.130
  Sex                                                    0.0566
    Male                                       8 (72.7)
    Female                                     3 (27.3)
  Weight, kg                     60.2 (11.6)             0.006
  Body mass index, kg/[m.sup.2]  23.7 (3.7)              0.015
Medical history
  Heart failure                                2 (18.2)  0.3171
  Oncologic pathology                          2 (18.2)  0.2897
Basal situation
  Dyspnea                                                0.00002
    Grade 0                                    0
    Grade 1                                    0
    Grade 2                                    0
    Grade 3                                    3 (27.3)
    Grade 4                                    8 (72.7)
  Cor pulmonale                                6 (54.5)  0.2721
  Pa[O.sub.2], mm Hg             63.2 (5)                0.094
  PaC[O.sub.2], mm Hg            45.8 (5)                0.211
  FE[V.sub.1], %                 32.4 (5)                0.017
  FE[V.sub.1]/FVC                36.0 (7)                0.001
Habitual therapy
  Long-term oxygen therapy                     8 (72.7)  0.0006
  Continuous oral corticoids                   5 (45.5)  0.0396
Admission data
  Changes in expectoration                     8 (72.7)  0.0220
  Increased symptoms, d           7.3 (8.5)              0.034
  Respiratory rate               34.0 (11.1)             0.012
  Thoracic excursion, cm          1.3 (0.9)              0.018
  Level of consciousness                                 0.00002
  Alert                                        6 (54.5)
  Confusion                                    2 (18.2)
  Stupor                                       3 (27.3)
  Coma                                         0
  Pa[O.sub.2], mm Hg*            50.1 (12)               0.217
  PaC[O.sub.2], mm Hg*           50.6 (15)               0.328
  PEF, L/min                     78.5 (51)               0.0001

(a) PEF, peak expiratory flow.
*Inspired oxygen concentration, 21%.

Table 3. Validity and reliability of functional tests and peak
expiratory flow rate in relation to in-patient mortality indicating
maximum sensitivity cutoff points

                        FE[V.sub.1]              FE[V.sub.1]/FVC
                        [less than or equal to]  [less than or equal to]
                        42%                      47%

Sensitivity (95% CI)    100 (67.8-99.1)          100 (67.8-99.1)
Specificity (95% CI)     46.9 (40.9-52.9)         49.8 (43.8-55.8)
Positive predictive       6.96 (3.7-12.4)          7.33 (3.9-13.0)
  value (95% CI)
Negative predictive     100 (96.4-99.9)          100 (96.6-99.9)
  value (95% CI)
Overall value (95% CI)   48.9 (43.0-54.8)         51.7 (45.8-57.6)
Likelihood ratio          1.88 (1.6-2.1)           1.99 (1.7-2.2)
  (95% CI)

                        PEF [less than or equal to] 150 L/min

Sensitivity (95% CI)    100 (67.8-99.1)
Specificity (95% CI)     51.9 (45.9-57.9)
Positive predictive       7.63 (4.0-15.5)
  value (95% CI)
Negative predictive     100 (96.7-99.9)
  value (95% CI)
Overall value (95% CI)   53.8 (47.8-59.6)
Likelihood ratio          2.08 (1.8-2.3)
  (95% CI)

95% CI, 95% confidence interval.

Table 4. Multivariate analysis of mortality (a)

Variable                                  OR     95% CI      P

Long-term domiciliary oxygen therapy      12.46  2.14-72.40  0.0049
Body mass index, kg/[m.sup.2]              0.73  0.59-0.90   0.0034
Peak expiratory flow on admission, L/min   0.96  0.94-0.98   0.0003

(a) OR, odds ratio; 95% CI, 95% confidence interval.
*Analysis takes into account age, sex, weight, history of heart failure
or oncologic pathology, oral corticoid therapy, degree of dyspnea,
thoracic excursion value, FE[V.sub.1]%, FE[V.sub.1]/FVC, number of days
with increased symptoms, changes in expectoration, respiration rate,
and level of consciousness on admission.

Accepted October 6, 2004.

Please see Janet M. Shapiro's editorial on page 263 of this issue.

References

1. Higgins MW. Chronic airways disease in the United States: trends and determinants. Chest 1989;96(Suppl 3):S328-S334.

2. Mannino DM, Homa DM, Akinbami LJ, et al. Chronic obstructive pulmonary disease surveillance: United States, 1971-2000. MMWR Surveil Summ 2002;51(SS-6):1-16.

3. Comite Cientifico del Estudio IBERCOP. IBERCOP: valoracion de resultados. Arch Bronconeumol 1999;35(Suppl 3):S40-S43.

4. Ward MM, Javitz HS, Smith WM, et al. Direct medical cost of chronic obstructive pulmonary disease in USA. Respir Med 2000;94:1123-1129.

5. Speizer FE. The rise in chronic obstructive pulmonary disease mortality: overview and summary. Am Rev Respir Dis 1989;140(Suppl 3):S106-S107.

6. Chailleux E, Faouroux B, Binet F, et al. Predictors of survival in patients receiving domiciliary oxygen therapy or mechanical ventilation: a 10-year analysis of ANT ADIR observatory. Chest 1996;109:741-749.

7. Hodgkin JE. Prognosis in chronic obstructive pulmonary disease. Clin Chest Med 1990;11:555-569.

8. Schols AM, Slangen J, Volovics L. Weight loss is a reversible factor in the prognosis of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998;157:1791-1797.

9. Ashutosh K, Haldipur C, Boucher ML. Clinical and personality profiles and survival in patients with COPD. Chest 1997;111:95-99.

10. Vestbo J, Prescott E, Lange P, et al. Vital prognosis after hospitalization for COPD: a study of a random population sample. Respir Med 1998;92:772-776.

11. Incalzi RA, Fuso L, De Rosa M, et al. Electrocardiographic signs of chronic cor pulmonale: a negative prognostic finding in chronic obstructive pulmonary disease. Circulation 1999;99:1600-1605.

12. Oswald-Mammosser M, Weitzenblum E, Quoix E, et al. Prognostic factors in COPD patients receiving long-term oxygen therapy: importance of pulmonary artery pressure. Chest 1995;107:1193-1198.

13. Fuso L, Incalci RA, Pistelli R, et al. Predicting mortality of patients hospitalized for acutely exacerbated chronic obstructive pulmonary disease. Am J Med 1995;98:272-277.

14. Connors AF Jr, Dawson NV, Thomas C, et al. Outcomes following acute exacerbation of severe chronic obstructive lung disease: the SUPPORT investigators. Am J Respir Crit Care Med 1996;154:956-967.

15. Barbera JA, Peces-Barba G, Agusti AGN, et al. Guia clinica para el diagnostico y el tratamiento de la enfermedad pulmonar obstructiva cronica. Arch Bronconeumol 2001;37:297-316.

16. Pauwels RA, Buist AS, Calverley PMA, et al, GOLD Scientific Committee. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: National Heart, Lung and Blood Institute and World Health Organization Global Initiative for Chronic Obstructive Lung Disease (GOLD): executive summary. Am J Respir Crit Care Med 2001;163:1256-1276.

17. ATS Statement: Standards for the diagnosis and care of patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1995;152 (Suppl 5):S77-120.

18. Ferris BG. Epidemiology standardization project, II: recommend respiratory disease questionnaires for use with adults and children in epidemiological research. Am Rev Respir Dis 1978;118:7-53.

19. Mahler DA, Wells CK. Evaluation of clinical methods for rating dyspnea. Chest 1998;93:580-586.

20. Pratt PC. Role of conventional chest radiography in diagnosis and exclusion of emphysema. Am J Med 1987;82:998-1006.

21. Vandiviere HM. Pulmonary hypertension and cor pulmonale. South Med J 1993;86:2S7-2S10.

22. Siafakas NM, Vermeire P, Pride NB, et al. ERS-Consensus Statement: Optimal assessment and management of chronic obstructive pulmonary disease (COPD). Eur Respir J 1995;8:1398-1420.

23. American Thoracic Society. Standardization of spirometry and PEF: 1994 update. Am J Respir Crit Care Med 1995;152:1107-1136.

24. Roca J, Sanchis J, Agusti-Vidal A, et al. Spirometric reference values for a Mediterranean population. Bull Eur Physiopathol Respir 1986;22:217-224.

25. Vermeire P. The burden of chronic obstructive pulmonary disease. Respir Med 2002;96(Suppl C):S3-S10.

26. Poole PJ, Bagg B, Brodie SM, et al. Characteristics of patients admitted to hospital with chronic obstructive pulmonary disease. N Z Med J 1997;110:272-275.

27. Mushlin AI, Black ER, Connolly CA, et al. The necessary length of hospital stay for chronic pulmonary disease. JAMA 1991;266:80-83.

28. Neill AM, Epton MJ, Martin IR, et al. An audit of the assessment and management of patients admitted to Christchurch Hospital with chronic obstructive pulmonary disease. N Z Med J 1994;107:365-367.

29. Medical Research Council Working Party. Long-term domiciliary oxygen therapy in chronic hypoxic cor pulmonale complicating chronic bronchitis and emphysema. Lancet 1981;1:681-685.

30. Nocturnal Oxygen Therapy Trial Group. Continuous or nocturnal oxygen therapy in hypoxemic chronic obstructive lung disease. Ann Intern Med 1980;93:391-398.

31. Wilson DO, Rogers RM, Wright EC, et al. Body weight in chronic obstructive pulmonary disease: National Institutes of Health IPPB Trial. Am Rev Respir Dis 1989;139:1435-1438.

32. De Miguel J, Grau T, Izquierdo JL. Papel de la nutricion en la enfermedad pulmonar obstructiva cronica. Med Clin (Barc) 1998;110:307-316.

33. Murata GH, Lium DJ, Busby HK, et al. Precision and accuracy of self-measured peak expiratory flow rates in chronic obstructive pulmonary disease. South Med J 1998;91:919-924.

34. Paggiaro PL, Moscato G, Giannini D, et al. Relationship between peak expiratory flow (PEF) and FE[V.sub.1]. Eur Respir J 1997;10(Suppl 24):S39-S41.

35. Sherwood P. The relationship between peak expiratory flow and respiratory symptoms. Eur Respir J 1997;10(Suppl 24):S67-S68.

36. Murata GH, Kapsner CO, Lium DJ, et al. A multivariate model for predicting respiratory status in patients with chronic obstructive pulmonary disease. J Gen Intern Med 1998;13:462-468.

37. Emerman CL, Cydulka RK. Use of peak expiratory flow rate in emergency department evaluation of acute exacerbation of chronic obstructive pulmonary disease. Ann Emerg Med 1996;27:159-163.

38. Hansen EF, Vestbo J, Phanareth K, et al. Peak flow as predictor of overall mortality in asthma and chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163:690-693.

39. Burgueno MJ, Garcia-Bastos JL, Gonzalez-Buitrago JM. Las curvas ROC en la evaluation de las pruebas diagnosticas. Med Clin (Barc) 1995;104:661-670.

RELATED ARTICLE: Key Points

* Chronic obstructive pulmonary disease is a major cause of chronic morbidity and death throughout the world.

* Exacerbations of respiratory symptoms requiring medical intervention are important clinical events in chronic obstructive pulmonary disease, and the economic and social burden is extremely high.

* The risk of dying from an acute exacerbation of chronic obstructive pulmonary disease is closely related to the development of respiratory acidosis, the presence of significant comorbidities, and the need for ventilatory support.

* A peak expiratory flow below 150 L/min is a strong predictor of death in chronic obstructive pulmonary disease, and identifies patients at high risk that may need continuous monitoring and intensive treatment.

Fernando de la Iglesia, MD, PHD, Jose Luis Diaz, MD, Salvador Pita, MD, PHD, Ricardo Nicolas, MD, PHD, Vicente Ramos, MD, Carlos Pellicer, MD, and Fernando Diz-Lois, MD

From Unidad de Corta Estancia Medica (UCEM) and Unidad de Epidemiologia Clinica y Bioestadistica. Complexo Hospitalario Universitario Juan Canalejo, A Coruna, Spain.

Reprint requests to Dr. Fernando de la Iglesia, UCEM, [3.sup.a] planta, Complexo Hospitalario Juan Canalejo, Xubias de Arriba, 84, 15006, A Coruna. Spain. Email: [e-mail]