- Email: [email protected]
Corticosteroid Therapy and Relapse in Sarcoidosis
Limitations of ECG in Diagnosing Pulmonary Embolism
REFERENCES
To the Editor:
2
In the article "The ECG in Pulmonary Embolism" by Ferrari et al (March 1997), 1 the authors report their experience with the use of ECG in monitoring the workouts of patients with pulmonary embolism (PE). They found that the pattern of inverted T waves in the anterior precordial leads is the most frequent ECG sign in PE patients and the best index of massive PE, being associated with an angiographic Miller index >50% and a mean pulmonary arterial pressure >30 mm Hg. Other ECG signs are reported with variable frequ encies, often strikingly different from those reported in other papers, 2 - 4 eg, sinus tachycardia is rated just at 26%. Although this paper has th e undoubted me1it of refocusing the role of the ECG in the complex diagnostic process of PE, some limitations should be considered. First, the authors do not stress the most important role of ECG in PE, ie, the capability of raising the suspicion of this condition or reinforcing clinical suspicion. 3 To this end, most ECG abnormalities, even those totally nonspecific, such as sinus tachycardia or S-T depression, should deserve particular attention when no definite coronary heart disease or other cardiac disorder can be diagnosed. Indeed, such signs have a high prevalence (about 50%) in all the patients referred with the suspicion of PE , and their frequency remains high in patients with confirmed PE without underlying cardiopulmonary disease. 3 In addition, the presence of S-T wave depression is also associated with the severity of PE. 3 Therefore, the presence of the above two ECG signs, along with a few others (ie, late R wave in aVR, P-R displacement, S slurred and T inversion in right precordial leads, S1 Q3 T 3 ) that are less frequent in all the patients with PE suspicion but still are significantly more frequent in those with confirmed PE, should always raise the suspicion of PE and prompt the pertinent diagnostic workout. Secondly, the series of patients reported is affected by a remarkable selection bias: the PE population has an unusually high incidence of massive PE , as underlined in the editorial by Moser. 5 Indirectly, this is confirmed by the high frequency of McGinn-White sign, 50% in their patients, whereas it is just 16% in our series of 145 consecutive patients with confirmed PE. 3 Our patients came from various hospital departments (including th e surgical ones) and presented with a continuous spectrum of PE severity (the number of unperfused lung segments was normally distributed), the mean obstruction of pulmonary perfusion being around 50% of the totaP Finally, the authors do not state explicitly that the most important conclusion of their paper, that "the anterior subepicardial ischemic pattern is closely related to the initial severity of PE," had been previously reported,3 not just suspected. We had shown an association between the presence ofT wave inversion in right precordial leads and the numbe r of unperfused lung segments on perfusion lung scan (p<0.05) and that, also very importantly, this ECG sign was significantly less frequent at recovery from PE when the scan showed a significant perfusion improvement. 3 Perhaps, our data should have been reported in Table 1.
3
Stefano Petruzzelli, MD, PhD Antonio Palla, MD Carlo Giuntini MD, FCCP CNR Institute of Clinical Physiology Pisa, Italy
4 5
Ferrari E, Imbert A, Chevalier T, e t al. The ECG in pulmonary embolism: prospective value of negative T waves in precordialleads-80 case reports. Chest 1997; 111:537-43 UPET. Clinical and electrocardiographic observations. Circulation 1973; 47:60-65 Petruzzelli S, Palla A, Pieraccini F , e t al. Routine electrocardiography in screening for pulmonary embolism. Respiration 1986; 50:233-43 Manganelli D, Palla A, Donnamaria V, eta!. Clinical features of pulmonary embolism: doubts and unce1tainties. Chest 1995; l07:25S-32S Moser KM. New and old questions about acute venous thromboembolism . Chest 1997; 111:531-32
To the Editor:
Thank you for offering me an opportunity to respond to the letter from Petruzzelli et al, concerning our paper published in CHEST (March 1997). 1 As we repeated s everal tim es in our article, we did not attempt to study the diagnostic usefulness of ECG in pulmonary embolism (PE ) as much as its usefulness as a predictor of severity in PE that has already been diagnosed elsewhere. Similarly, the fact that the PE patients were selected from a cardiology ward does not constitute a "remarkable selection bias" in as far as this was the population that was targeted in the aim of the study. These patients in fact form ed a specific group and reflected a clinical reality. Last, regarding the study cited,2 I think that pulmonary angiography remains the "gold standard" for evaluating severity in cases of PE, and that a correlation between the ECG signs and scintigraphic signs cannot be as reliable a s when the reference adopted is the accepted standard. Emile Ferrari, MD Hopital Pasteur Nice, France
REFERENCES 1 F errari E, Imbe1t A, Chevalier T, e t a!. The ECG in pulmonary embolism: prospective value of negative T waves in precordial leads-80 case reports. Chest 1997; 111:537-43 2 Petruzzelli S, Palla A, Pieraccini F, eta!. Routine e le ctrocardiography in screening for pulmonary embolism. Respiration 1986; 50:233-43
Corticosteroid Therapy and Relapse in Sarcoidosis To the Editor:
Despite reading this very interesting article by Gottlieb et al (March 1997)1 with great care four times, I (and perhaps others as well) remain puzzled by a number of points that I hope th e authors can resolve. l. The proportion (65%) of patients who were judged to require corticosteroid therapy (CST) because of symptoms is unusually high. In a population-based study of this disease, 2 in which African-Americans constituted 27%, only 25% received CST despite the employment ofless stringent indications (other than symptoms) in some instances (eg, failure to resolve radiographic abnormalities or physiologic or radiographic progression). What were the sources for these the patients? Was there a tertiary referral screen such that the population under study might have been adversely selected? 2. The ratio of females to males in the Caucasian subjects was CHEST I 113 I 2 I FEBRUARY, 1998
559
nearly 1:1; among Af1ican-Americans it was 3: 1. Do the authors have an explanation for the disproportionate involvement of African-American fe male.;? 3. A mortality of 10% attributable to sarcoidosis was estim ated in an earlier study from this institution. 3 Is it correct to infer from the statement "no patients were considered 'lost' to follow-up »! that there were no deaths in 3,645 at-risk patient years of th e study? If so, how do the authors account for the change in outcome? 4. Although the group composition was well matched demographically, no information was provided concerning group allocation by stage or disease duration, both of which substantially influence prognosis. Can the authors provide this information? 5. The authors employ the terms "relapse," meaning recrudescence of symptoms, and "re mission," meaning resolution of symptoms. \Vhat do these te rms signify in patients in the "Spontaneous Group" who have t1ivial or no symptoms? Why, for example, would one choose to assign the term "relapse," rather than "progression," to a previously asymptomatic patient whose disease was worsen ing? 6. I employ th e same usage as Badrinas et al,4 who distinguish "recurrence," meaning the reappearance of sarcoidosis follmving complete spontaneous clinical and radiographic resolution, from "relapse," meaning recrudescence of latent disease following partial or co mplete CST-induced remission. The form er appears to be a rare event, identified in 1 of 210 patients reported by Rom er,5 and in 0 of 86 patients reported b y Re ichand Johnson.2 Do the authors agree with this distinction? Finally, I should like to make four comme nts: 1. There isan inconsequential e rror in th e computed p ercentage (25%) in th e sentence "In the British Thoracic Society study, 5 of 25=25%." It should be "S of 25=20% ." 2. The inte resting speculation that th e authors provide to account for the superior course expe rie nced by the untreated patients-"corticosteroid treatment. .. co ntributed to the propensity for relapse" 1- is supported among persons \vith recently di agnosed sarcoidosis not only by th e cited r eports of Eule e t al 6 and Izumi 7 but also by an earlier study from their institution. In that study, a randomized double-blind trial of prednisone, a course of 15 mg daily for 3 months was administered to 83 patients; 21 of 46 pe rsons with stage II or III disease received ean observation prednisone; 25 were untreated controls. After a m period of 5.3 years, clinical, spirometric, and radiographic evaluation showed a more favorable outcome in the stage II and III untreated group (treated/untreated ): d efinite improvement, 48%/ 44%; progression, 38%/16%; persistent radiographi c abnormality, 71 %/60%. Th ese diffe rences did not ach ieve a 0.1 level of significances Two patients, whose allocation was not specified, died of sarcoidosis. Can the authors provide their allocati on? The outco me of these three studies may not be generalizable to the constitu en t population of this report whose disease was, for the most part, chroni c. The recently published British Thoracic Society rep01i9 (ci ted b y the authors ), which demonstrated a small long-term advantage in the treated cohort, involved pati ents similar in age and th e refore probably similar in disease duration. 3. The authors state : "The purpose of highlighting these two groups [induced and spontaneous] was to explore the possible impact of corticosteroid therapy by co mparing two groups of stable, asymptomatic patients with sarcoidosis, only one of which had been treated"1 [i talics added]. This appears to be at variance with the earlier statement that the induced group was treated for "compelling symptoms," and that the spontaneous group had either "asymptomatic radiographic abnormalities or initial symptoms that were not judged of sufficient severity to warrant [co rticosteroid] treatment. "1 560
4. The author's hypothesis that severity of symptoms (and consequent need for CST) accurately predicts an adverse course is supp01ied by Kolek,10 who observed the converse, a highly favorable course among 1,500 patients with sarcoidosis who were either asymptomatic or had trivial symptoms. Whichever hypothesis one accepts as the most probable-that severe symptoms are an accurate marker of prognosis or that CST adversely affects the course of th e disease in some individuals-and the two are not mutually exclusive, the lesson seems clear: CST should be e mployed with restraint, if at all, in persons with trivial or no symptoms. What do the authors advocate for those patients who are either symptomatic or who are asymptomatic but demonstrate clear evidence of radiographi c or physiologic progression? Jerome M. Reich, MD, FCCP Center f or Health Research Portland, Oregon REFERENCES 1 Gottlieb JE, Israel HL, Steiner RM , et !a. Outcome in sarcoidosis: the relationship of relapse to corticosteroid therapy. Chest 1997; 111:623-31 2 R eich JM , Johnson RE. Course and prognosis of sarcoidosis in a nonreferral setting: analysis of 86 patients observed for 10 years. Am J M ed 198.5; 78:61-67 3 Sones M, Israel HL. Course and prognosis of sarcoidosis. Am J M ed 1960; 29:84-93 4 Badrinas F , Mana J, Fernandez-Nogues F. Recurrent sarcoidosis. Sarcoidosis 1989; 6:63-64 5 Rom er FK. Presentation of sarcoidosis and outcome of pulmonary changes. D an Med Bull1982; 29:27-32 6 Eule II, Weinecke A, Roth I. The possible influence of corticosteroid therapy on the natural course of pulmonary sarcoidosis. Ann NY Acad Sci 1986; 465:69.5-701 7 Izumi T. Are corticosteroids harmful to sarcoidosis? Sarcoidosis 1994; 11:119-22 8 Israel HL, Fouts DW, Beggs RA. A controlled trial of prednisone treatment of sarcoidosis. Am Rev Respir Dis 1973; 107:609-14 9 Gibson GJ, Prescott RT, Muers MF, et la. British Thoracic Society sarcoidosis study: effects of long te rm corticosteroid treatment. Thorax 1996; 51:238-47 10 Kolek V. Epidemiologic study on sarcoidosis in Moravia and Silesia. Sarcoidosis 1994; 11:110-12 To the Editor: Thank you for the opportunity to reply to the letter from Dr. Reich. Our proportion of patients judged to require corticosteroid pattern to a tertiary r th erapy (65%) reflects, we believe, a eferral institution and to a r ne owned clinician (H. L. Israel). Our impression is that many of these patients presented with difficult, puzzling, or severe cases of sarcoidosis. vVe do not believe that this represents an adverse selection of patients; it represents a non-random selection. In addition, th e decision to treat was made physician, reflecting an identified source of bias. s by a ingle Howeve r, we believe the advantage of the single physician decision point is that it resulted in a consistent bias. The ratio of females to males present in this study was observed from referral patterns among patients who were seen at multiple visits over a 4-year period. This observation allows for hypotheses ranging from matters of disease prevalence to cultural values to selection bias; we leave the generation of these or other hypotheses to the re ader. vVe would caution Dr. Reich or any other reader against making inferences from our study about prevalence or incidence. Our patients were selected in part by demonstrated interaction \\~th their physician on at least two occasions over the 4-year pe Jiod. Therefore, patients who had been seen previously but who had died were, by definition, not included in this study. Communications to the Editor
REFERENCES
To the Editor:
2
In the article "The ECG in Pulmonary Embolism" by Ferrari et al (March 1997), 1 the authors report their experience with the use of ECG in monitoring the workouts of patients with pulmonary embolism (PE). They found that the pattern of inverted T waves in the anterior precordial leads is the most frequent ECG sign in PE patients and the best index of massive PE, being associated with an angiographic Miller index >50% and a mean pulmonary arterial pressure >30 mm Hg. Other ECG signs are reported with variable frequ encies, often strikingly different from those reported in other papers, 2 - 4 eg, sinus tachycardia is rated just at 26%. Although this paper has th e undoubted me1it of refocusing the role of the ECG in the complex diagnostic process of PE, some limitations should be considered. First, the authors do not stress the most important role of ECG in PE, ie, the capability of raising the suspicion of this condition or reinforcing clinical suspicion. 3 To this end, most ECG abnormalities, even those totally nonspecific, such as sinus tachycardia or S-T depression, should deserve particular attention when no definite coronary heart disease or other cardiac disorder can be diagnosed. Indeed, such signs have a high prevalence (about 50%) in all the patients referred with the suspicion of PE , and their frequency remains high in patients with confirmed PE without underlying cardiopulmonary disease. 3 In addition, the presence of S-T wave depression is also associated with the severity of PE. 3 Therefore, the presence of the above two ECG signs, along with a few others (ie, late R wave in aVR, P-R displacement, S slurred and T inversion in right precordial leads, S1 Q3 T 3 ) that are less frequent in all the patients with PE suspicion but still are significantly more frequent in those with confirmed PE, should always raise the suspicion of PE and prompt the pertinent diagnostic workout. Secondly, the series of patients reported is affected by a remarkable selection bias: the PE population has an unusually high incidence of massive PE , as underlined in the editorial by Moser. 5 Indirectly, this is confirmed by the high frequency of McGinn-White sign, 50% in their patients, whereas it is just 16% in our series of 145 consecutive patients with confirmed PE. 3 Our patients came from various hospital departments (including th e surgical ones) and presented with a continuous spectrum of PE severity (the number of unperfused lung segments was normally distributed), the mean obstruction of pulmonary perfusion being around 50% of the totaP Finally, the authors do not state explicitly that the most important conclusion of their paper, that "the anterior subepicardial ischemic pattern is closely related to the initial severity of PE," had been previously reported,3 not just suspected. We had shown an association between the presence ofT wave inversion in right precordial leads and the numbe r of unperfused lung segments on perfusion lung scan (p<0.05) and that, also very importantly, this ECG sign was significantly less frequent at recovery from PE when the scan showed a significant perfusion improvement. 3 Perhaps, our data should have been reported in Table 1.
3
Stefano Petruzzelli, MD, PhD Antonio Palla, MD Carlo Giuntini MD, FCCP CNR Institute of Clinical Physiology Pisa, Italy
4 5
Ferrari E, Imbert A, Chevalier T, e t al. The ECG in pulmonary embolism: prospective value of negative T waves in precordialleads-80 case reports. Chest 1997; 111:537-43 UPET. Clinical and electrocardiographic observations. Circulation 1973; 47:60-65 Petruzzelli S, Palla A, Pieraccini F , e t al. Routine electrocardiography in screening for pulmonary embolism. Respiration 1986; 50:233-43 Manganelli D, Palla A, Donnamaria V, eta!. Clinical features of pulmonary embolism: doubts and unce1tainties. Chest 1995; l07:25S-32S Moser KM. New and old questions about acute venous thromboembolism . Chest 1997; 111:531-32
To the Editor:
Thank you for offering me an opportunity to respond to the letter from Petruzzelli et al, concerning our paper published in CHEST (March 1997). 1 As we repeated s everal tim es in our article, we did not attempt to study the diagnostic usefulness of ECG in pulmonary embolism (PE ) as much as its usefulness as a predictor of severity in PE that has already been diagnosed elsewhere. Similarly, the fact that the PE patients were selected from a cardiology ward does not constitute a "remarkable selection bias" in as far as this was the population that was targeted in the aim of the study. These patients in fact form ed a specific group and reflected a clinical reality. Last, regarding the study cited,2 I think that pulmonary angiography remains the "gold standard" for evaluating severity in cases of PE, and that a correlation between the ECG signs and scintigraphic signs cannot be as reliable a s when the reference adopted is the accepted standard. Emile Ferrari, MD Hopital Pasteur Nice, France
REFERENCES 1 F errari E, Imbe1t A, Chevalier T, e t a!. The ECG in pulmonary embolism: prospective value of negative T waves in precordial leads-80 case reports. Chest 1997; 111:537-43 2 Petruzzelli S, Palla A, Pieraccini F, eta!. Routine e le ctrocardiography in screening for pulmonary embolism. Respiration 1986; 50:233-43
Corticosteroid Therapy and Relapse in Sarcoidosis To the Editor:
Despite reading this very interesting article by Gottlieb et al (March 1997)1 with great care four times, I (and perhaps others as well) remain puzzled by a number of points that I hope th e authors can resolve. l. The proportion (65%) of patients who were judged to require corticosteroid therapy (CST) because of symptoms is unusually high. In a population-based study of this disease, 2 in which African-Americans constituted 27%, only 25% received CST despite the employment ofless stringent indications (other than symptoms) in some instances (eg, failure to resolve radiographic abnormalities or physiologic or radiographic progression). What were the sources for these the patients? Was there a tertiary referral screen such that the population under study might have been adversely selected? 2. The ratio of females to males in the Caucasian subjects was CHEST I 113 I 2 I FEBRUARY, 1998
559
nearly 1:1; among Af1ican-Americans it was 3: 1. Do the authors have an explanation for the disproportionate involvement of African-American fe male.;? 3. A mortality of 10% attributable to sarcoidosis was estim ated in an earlier study from this institution. 3 Is it correct to infer from the statement "no patients were considered 'lost' to follow-up »! that there were no deaths in 3,645 at-risk patient years of th e study? If so, how do the authors account for the change in outcome? 4. Although the group composition was well matched demographically, no information was provided concerning group allocation by stage or disease duration, both of which substantially influence prognosis. Can the authors provide this information? 5. The authors employ the terms "relapse," meaning recrudescence of symptoms, and "re mission," meaning resolution of symptoms. \Vhat do these te rms signify in patients in the "Spontaneous Group" who have t1ivial or no symptoms? Why, for example, would one choose to assign the term "relapse," rather than "progression," to a previously asymptomatic patient whose disease was worsen ing? 6. I employ th e same usage as Badrinas et al,4 who distinguish "recurrence," meaning the reappearance of sarcoidosis follmving complete spontaneous clinical and radiographic resolution, from "relapse," meaning recrudescence of latent disease following partial or co mplete CST-induced remission. The form er appears to be a rare event, identified in 1 of 210 patients reported by Rom er,5 and in 0 of 86 patients reported b y Re ichand Johnson.2 Do the authors agree with this distinction? Finally, I should like to make four comme nts: 1. There isan inconsequential e rror in th e computed p ercentage (25%) in th e sentence "In the British Thoracic Society study, 5 of 25=25%." It should be "S of 25=20% ." 2. The inte resting speculation that th e authors provide to account for the superior course expe rie nced by the untreated patients-"corticosteroid treatment. .. co ntributed to the propensity for relapse" 1- is supported among persons \vith recently di agnosed sarcoidosis not only by th e cited r eports of Eule e t al 6 and Izumi 7 but also by an earlier study from their institution. In that study, a randomized double-blind trial of prednisone, a course of 15 mg daily for 3 months was administered to 83 patients; 21 of 46 pe rsons with stage II or III disease received ean observation prednisone; 25 were untreated controls. After a m period of 5.3 years, clinical, spirometric, and radiographic evaluation showed a more favorable outcome in the stage II and III untreated group (treated/untreated ): d efinite improvement, 48%/ 44%; progression, 38%/16%; persistent radiographi c abnormality, 71 %/60%. Th ese diffe rences did not ach ieve a 0.1 level of significances Two patients, whose allocation was not specified, died of sarcoidosis. Can the authors provide their allocati on? The outco me of these three studies may not be generalizable to the constitu en t population of this report whose disease was, for the most part, chroni c. The recently published British Thoracic Society rep01i9 (ci ted b y the authors ), which demonstrated a small long-term advantage in the treated cohort, involved pati ents similar in age and th e refore probably similar in disease duration. 3. The authors state : "The purpose of highlighting these two groups [induced and spontaneous] was to explore the possible impact of corticosteroid therapy by co mparing two groups of stable, asymptomatic patients with sarcoidosis, only one of which had been treated"1 [i talics added]. This appears to be at variance with the earlier statement that the induced group was treated for "compelling symptoms," and that the spontaneous group had either "asymptomatic radiographic abnormalities or initial symptoms that were not judged of sufficient severity to warrant [co rticosteroid] treatment. "1 560
4. The author's hypothesis that severity of symptoms (and consequent need for CST) accurately predicts an adverse course is supp01ied by Kolek,10 who observed the converse, a highly favorable course among 1,500 patients with sarcoidosis who were either asymptomatic or had trivial symptoms. Whichever hypothesis one accepts as the most probable-that severe symptoms are an accurate marker of prognosis or that CST adversely affects the course of th e disease in some individuals-and the two are not mutually exclusive, the lesson seems clear: CST should be e mployed with restraint, if at all, in persons with trivial or no symptoms. What do the authors advocate for those patients who are either symptomatic or who are asymptomatic but demonstrate clear evidence of radiographi c or physiologic progression? Jerome M. Reich, MD, FCCP Center f or Health Research Portland, Oregon REFERENCES 1 Gottlieb JE, Israel HL, Steiner RM , et !a. Outcome in sarcoidosis: the relationship of relapse to corticosteroid therapy. Chest 1997; 111:623-31 2 R eich JM , Johnson RE. Course and prognosis of sarcoidosis in a nonreferral setting: analysis of 86 patients observed for 10 years. Am J M ed 198.5; 78:61-67 3 Sones M, Israel HL. Course and prognosis of sarcoidosis. Am J M ed 1960; 29:84-93 4 Badrinas F , Mana J, Fernandez-Nogues F. Recurrent sarcoidosis. Sarcoidosis 1989; 6:63-64 5 Rom er FK. Presentation of sarcoidosis and outcome of pulmonary changes. D an Med Bull1982; 29:27-32 6 Eule II, Weinecke A, Roth I. The possible influence of corticosteroid therapy on the natural course of pulmonary sarcoidosis. Ann NY Acad Sci 1986; 465:69.5-701 7 Izumi T. Are corticosteroids harmful to sarcoidosis? Sarcoidosis 1994; 11:119-22 8 Israel HL, Fouts DW, Beggs RA. A controlled trial of prednisone treatment of sarcoidosis. Am Rev Respir Dis 1973; 107:609-14 9 Gibson GJ, Prescott RT, Muers MF, et la. British Thoracic Society sarcoidosis study: effects of long te rm corticosteroid treatment. Thorax 1996; 51:238-47 10 Kolek V. Epidemiologic study on sarcoidosis in Moravia and Silesia. Sarcoidosis 1994; 11:110-12 To the Editor: Thank you for the opportunity to reply to the letter from Dr. Reich. Our proportion of patients judged to require corticosteroid pattern to a tertiary r th erapy (65%) reflects, we believe, a eferral institution and to a r ne owned clinician (H. L. Israel). Our impression is that many of these patients presented with difficult, puzzling, or severe cases of sarcoidosis. vVe do not believe that this represents an adverse selection of patients; it represents a non-random selection. In addition, th e decision to treat was made physician, reflecting an identified source of bias. s by a ingle Howeve r, we believe the advantage of the single physician decision point is that it resulted in a consistent bias. The ratio of females to males present in this study was observed from referral patterns among patients who were seen at multiple visits over a 4-year period. This observation allows for hypotheses ranging from matters of disease prevalence to cultural values to selection bias; we leave the generation of these or other hypotheses to the re ader. vVe would caution Dr. Reich or any other reader against making inferences from our study about prevalence or incidence. Our patients were selected in part by demonstrated interaction \\~th their physician on at least two occasions over the 4-year pe Jiod. Therefore, patients who had been seen previously but who had died were, by definition, not included in this study. Communications to the Editor