Is the More (Intricate) the Better? To the Editor: Toraldo and colleagues (December 2005) 1 are to be congrat- ulated on their endeavor to identify daytime variables that are predictive of nocturnal desaturation in COPD patients, a field in which many attempts have failed. However, we think that some points are worth addressing. First, despite the plethora of data presented by Toraldo et al, 1 it does not greatly contribute to the evidence base of numerous previous articles that the probability for nonapneic desaturation is proportional to the severity of blood gas disturbances and lung function impairment. Second, we think that an analysis of the possible usefulness of the oxyhemoglobin dissociation curve 2 (ODC) as a predictive tool for nonapneic desaturation in COPD patients could have been included in the discussion. A physiologic decrease of Pao 2 during sleep in healthy subjects does not result in significant desatura- tion because of the plateau in this section of the lung. The decrease of Pao 2 in COPD patients frequently combined with an increase in Paco 2 and a decrease in pH, as well as with some specific changes in the biochemistry of hemoglobin (ie, an increase in 2,3-diphosphoglycerate and Po 2 corresponding to 50% saturation of hemoglobin) causes a rightward shift of ODC and a displacement of the desaturation point to the steeper slope of the ODC. 3 All of those conditions are prerequisites for significant nocturnal desaturation. We believe that the so-called capacitance coefficients that are a measure of the slope of different parts of the ODC 3 and their “desaturation capacity” may be a promising avenue for the more accurate prediction of the nocturnal desaturation in COPD patients. Having in mind the major clinical implications of significant nocturnal desaturation, we believe that the authors should com- ment on these issues. Stefan Kostianev, MD, DMSc Blagoi Marinov, MD, PhD Dimitar Iluchev, MD, DMSc Medical University of Plovdiv Plovdiv, Bulgaria The authors hereby declare that they have no conflicts of interest to disclose. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. org/misc/reprints.shtml). Correspondence to: Stefan Kostianev, MD, DMSc, Department of Pathophysiology, Medical University of Plovdiv, 15A Vassil Aprilov Blvd, 4002 Plovdiv, Bulgaria; e-mail: s_kostianev@ mail.orbitel.bg DOI: 10.1378/chest.130.4.1275 References 1 Toraldo DM, Nicolardi G, De Nuccio F, et al. Pattern of variables describing desaturator COPD patients, as revealed by cluster analysis. Chest 2005; 128:3828 –3837 2 Siggaard-Andersen O, Siggaard-Andersen M. The oxygen status algorithm: a computer program for calculating and displaying pH and blood gas data. Scand J Clin Lab Invest Suppl 1990; 203:29 – 45 3 Kostianev S, Iluchev D, Ivanova M. New parameters for evaluation of oxygen transport on the dissociation curve of human blood. Folia Med (Plovdiv) 1994; 36(1):5–12 To the Editor: We thank Dr. Kostianev et al for contribution of their knowl- edge in the field addressed by our article (December 2005). 1 We agree that independent predictors of nighttime desaturation in COPD patients useful for clinical applications are still being discussed. Some tests have been proposed: (1) low daytime Pao 2 values 2 ; (2) reduced sensitivity of respiratory centers to chemical stimuli 3 ; (3) serious alteration of pulmonary function test results; (4) a clinical frame of “blue bloaters”; (5) desaturation after physical effort; (6) measurement of carbon dioxide ventilatory response; and (7) awake daytime arterial oxygen saturation (Sao 2 ) values. Our study used for the first time a cluster analysis in such a prediction. The cluster analysis involves grouping similar objects into distinct, mutually exclusive subsets referred to as clusters. The elements within a cluster have a high degree of “natural association” among themselves, while the clusters are “relatively distinct” from one another. A cluster analysis method, therefore, can be simply defined as a procedure to classify data already used in clinical medicine for patients or patient data classification. Our data showed that i-cluster analysis was able to detect populations among both “desatutator” and “nondesaturator” COPD patients; desatutator patients were identified not by the percentage of total recording time (TRT) spent in bed with arterial oxygen saturation (Sao 2 )  90% alone, but rather by a pattern of percentage of TRT spent in bed with Sao 2  90% and mean pulmonary artery pressure and Paco 2 values, the two latter also being predictors of nocturnal desaturation severity. More- over, cluster analysis was able to identify subgroups of both desatutator and nondesaturator patients with varying degrees of illness. Kostianev et al 4 identified new biochemical parameters to evaluate oxygen transport using the dissociation curve of human hemoglobin. Siggaard-Andersen and Siggaard-Anderson 5 pro- posed a computer program for calculating and displaying pH and blood gas data that can be used to predict the oxygen status or acid-base status. Both methods are not simply to be used in clinical medicine for classification of patients. Our contribution is a further clinical/practical approach to clarify such a field opening to further studies on the sleep lung ventilation and oxygen saturation, the chemical control of respiratory function and on response to hypoxic and hypercapnic awake stimuli. Domenico Toraldo, MD Giuseppe Nicolardi, MD, PhD Francesco De Nuccio, PhD Nicolino Ambrosino, MD, FCCP A. Galateo Lung Disease Hospital Leece, Italy CHEST Correspondence www.chestjournal.org CHEST / 130 / 4 / OCTOBER, 2006 1275