Untreated Patients With Pulmonary Embolism

Untreated Patients With Pulmonary Embolism* Outcome, Clinical, and Laboratory Assessment Paul D . Stein, MD , FCCP; j erald W . Henry , MS; and Bruce Relyea, MD

Purpose: The purpose of this investigation was to evaluate the clinical characteristics of acute pulmonary embolism (PE) among patients with PE who did not receive treatment to assess how the natural course of untreated PE relates to its severity. Methods: Data are from the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED). Among patients with PE diagnosed either by pulmonary angiography or at autopsy, 376 received treatment and 24 escaped treatment in the hospital and were not prescribed anticoagulant therapy at the time of hospital discharge. Among these untreated patients, however, four received anticoagulant therapy during the first 3 months of follow-up. We report the 3-month course of PE in the remaining 20 untreated patients. Results: Among untreated patients, 1 of 20 (5.0% ) died from the effects of the original PE and possibly some contribution from recurrent PE. Nonfatal recurrent PE occurred in 1 of 20 (5.0% ). All untreated patients, 20 of 20 (100% ), had less than three mismatched segmental

perfusion defects compared with 227 of 376 (60% ) treated patients (p<0.001). Pulmonary arteries showed thromboemboli in segmental or smaller arteries in 16 of 19 (84% ) untreated patients compared with 132 of 362 (36% ) treated patients (p<0.001). Untreated and treated patients, when grouped according to the size of the ventilation/ perfusion lung scan defect or size of vessels involved on the pulmonary angiogram, showed no statistically significant difference in the frequency of fatal PE or nonfatal recurrent PE. Conclusion: Mild untreated PE carries a lower immediate mortality and lower mortality from recurrent PE than overt PE described in prior decades. (Chest 1995; 107:931-35)

The natural history of untreated pulmonary embolism (PE) is known primarily from investigations of clinically obvious and severe PE, diagnosed in an era before pulmonary angiography, and before the universal use of anticoagulants for this disorder.l· 2 Carson and associates,3 in an investigation of the clinical cause of acute PE, observed that in the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED), 24 patients escaped treatment in the hospital. This resulted primarily from a postdischarge interpretation of the pulmonary angiograms by a c ne tral consensus panel of expert readers, who , at leisure and without the exigencies of urgent care, found PE, contrary to the opinion of the local radiologists. Only one of these untreated patients described by Carson and associates3 died of PE. An indepth evaluation of these patients , their clinical characteristics, and the severity of their PE may

permit an assessment of why the death rate in untreated patients with acute PE was so much lower than previously has been reported.

*From the Henry Ford Heart and Vascular Institute, Detroit. Manuscript received May 16, 1994; revision accepted August 19. Reprint requests: Dr. Stein, Henry Ford Heart and Vascular Institute, New Cen ter Pavillion , Rm 1107, 2921 W Grand Blvd, Detroit, MI 48202

PE=pulmonary embolism; PIOPED=Prospective Investigation of Pulmonary Embolism Diagnosis

Key words: clinical diagnosis; deep venous thrombosis; noninvasive tests; pulmonary embolism; thromboembolism

M ETHODS

The diagnosis o fPE in all patients was Data are from PI made by pulmonary angiography o r au topsy. The present evalndomized t o obligatory anuation includes patients w ho were ra giography if they had an abnormal ventilation / perfusion lung scan, as reported in the p rim ary PIOPED report. 4 The present evaluation also includes patients in the arm of PI OPED in which for angiography only a t the request of the r patients were eferred attending physician. This arm of PI OPED was not included in the primary PIOPED report. Twenty-four patients in PIOPED escaped treatment in the hospital and received n oprescription f or anticoagulant therapy a t the tim e of hospital discharge. Am ong these, four were begun on a regimen of anticoagulant therapy during the first month after hospital discharge and one was begun on a regimen of anticoagulant therapy between the third and sixth month following hospital discharge. We evaluated the 20 patients who re ceived no trea tment for PE during the first 3 months of follow-up. The circumstances involving no therapy in these 20 patients are as follows: 19 had pulmonary angiograms interpreted as showing no PE by the local radiologist, but the interpretation of n o PE was reversed in 18 some tim e late r following reevaluation b y the cenogram readers and the diagnosis o f no PE was tral panel of angi OPED.4

CHEST I 107 I 4 I APRIL, 1995

931

Table 1-Age, Duration of Survival During 3-Month Follow-up Period, and Cause of Death Among 20 Untreated

Patients With Pulmonary Embolism

Age, yr

Duration of Survival, d

Autopsy Performed

Pulmonary Embolism Present at Autopsy

Pulmonary Embolism Cause of Death

Imm ediate Cause of Death

Cnderlying Cause of Death

75 63 33

4 6

No Yes Yes

Yes Yes

No No Yes

Mesenteric infarction Aspiration pneumonia Pulmonary embolism

64 67

20 20

Acute myocardial infarction Malignant neoplasm Primary pulmonary hypertension

1\o No

No No

Carcinoma of lung Septicemia

79 73 87

26 48 53

No No No

No No No

Pneumonia Septicemia Septicemia

reversed in 1 by the outcome classification committee because PE was found at autopsy 6 days after the angiogram. One patient had no angiogram, being in the arm of PIOPED that allowed angiography at the physician's request; the patient died of unrelated causes 4 days later and autopsy showed small peripheral PE. Techniques of recruiting, physical examination, ventilation/ perfusion lung scans, pulmonary angiography, and follow-up were reported in PIOPED 4 Outcome assessment, treatment, and follow-up procedures were described by Carson and associatesa Death and the presence of recurrent PE were assessed by an outcome-classification committee for 1 year after entry into the study. Cause of death was established after review of all available clinical information, findings at autopsy (if performed), and the patient's death certificate. Patients were classified as having died of PE if, in the opinion of the outcome classification committee, this condition was the immediate or underlying cause of death. Diagnostic studies for recurrent PE were ordered by the attending physician on the basis of clinical suspicion and were not part of the study protocol. Recurrent PE was determined by the outcome classification committee on the basis of recent PE at autopsy of patients who survived the original PE, by angiographic demonstration of new thromboemboli on follow-up, by the development of new perfusion lung scan defects consistent with a high probability of PE, or by a convincing clinical presentation. Among the untreated patients, three died in the hospital during the first week and five died during the first 3 months following hospital discharge (Table 1). Primary diagnoses of untreated patients who died are shown in Table l. With the exception of a patient with primary pulmonary hypertension who died of PE, all deaths were expected on the basis of unrelated disease. Patients who died and did not undergo autopsy did not have clinical manifestations suggestive of recurrent PE. Among patients with PE in the two arms of PIOPED, 376 received therapy that included full-dose heparin usually followed by oral anticoagulants or continuing full-dose heparin, inferior vena cava interruption, thrombolytic therapy, embolectomy, and in some instances, only antiplatelet agents or low-dose heparin. The four patients who received antiocoagulant therapy starting some time during the first month of follow-up were excluded. Among 376 treated patients, 297 were treated with full-dose anticoagulants, but not thrombolytic therapy , inferior vena cava interruption, or embolectomy, although some received anti platelet agents in addition. Among these 297 patients, 12 received full-dose heparin therapy but died prior to hospital discharge, 282 received full-dose heparin therapy followed by oral anticoagulants, and 3 received full-dose heparin therapy both in the hospital and after discharge.

932

Coronary heart disease, left and right ventricular failure COPD Carcinoma of colon Coronary heart disease, myocardial infarction

We compared the clinical characteristics, ventilation /perfusion lung scans, and pulmonary arteriograms among untreated and treated patients. Data on the history and physical examination, chest radiograph, pulmonary artery mean pressure, pulmonary arteriograms, and ventilation/ perfusion lung scans were evaluated. One segmental equivalent perfusion defect on a ventilation/perfusion lung scan was defined as one large segmental perfusion defect (>75% of a segment) or two moderate-sized (25 to 75% of a segment) segmental perfusion defects in a region without ventilation or chest radiographic abnormalities or substantially larger than either matching ventilation or chest radiographic abnormalities 4 The partial pressure of oxygen in arterial blood (Pa0 2 ), while breathing room air, was measured in 16 untreated patients and 272 treated patients. The alveolar arterial oxygen gradient among patients with Pa0 2 measured while breathing room air was calculated as described previously. 5

Statistical Methods A x2 with Yates' correction was used to compare the prevalence of symptoms and characteristics of laboratory tests. Comparisons of continuous variable means were made with Student's t test. Data are reported as mean± SD. RESULTS

Untreated Patients Death and Recurrent Pulmonary Embolism: Among 20 patients with PE who were untreated , 8 died (Table 1). Only one of eight deaths was due to PE. That patient, a 33-year-old woman, had underlying primary pulmonary hypertension with right ventricular failure. Organized and fresh pulmonary emboli were shown at autopsy 6 days after a pulmonary angiogram that failed to show PE. Whether this death resulted from the original PE or recurrent PE is uncertain . The frequency of death due to untreated PE was 1 of 20 (5.0%). One patient died 4 days after a ventilation/ perfusion lung scan interpreted as intermediate probability for PE was obtained. The patient died because of aspiration pneumonia following hysterectomy for endometrial carcinoma. Multiple small thromboemboli in peripheral branches were observed at autopsy. Outcome Assessment of Untreated PE Patients (Stein, Henry, Relyea)

Table 2-Clinical Findings Among Untreated and Treated Patients*

Dyspnea Pleuritic pain Hemoptysis Tachypnea (2:20/ min) Rales Tachycardia (> 100/ min ) Accentuated P2 Atelectasis/ parenchymal abnormality (radiograph) Pleural effusion (radiograph) Elevated diaphragm (radiograph) ST segment or T wave (ECG)

Untreated No. (%)

Treated No. (% )

p Value

12/ 20 (60) 6/ 20 (30) 2/ 20 (10) 12/ 20 (60) 12/ 20 (60) 7/ 20 (35) 5/ 20 (25) 14/ 20 (70)

300/ 376 (80) 227 / 376 (60) 61 / 376 (16) 278/ 376 (74) 208/ 376 (55) 114/ 376 (30) 86/ 376 (23) 258/ 376 (69)

NS <0.02 NS NS NS NS NS NS

8/ 20 (40) 6/ 20 (30)

180/ 376 (48) 103/ 376 (27 )

NS NS

11 / 18 (61)

154/ 299 (52)

NS

*P2=pulmonary component of second sound; V Q / =ventilatiori/ perfusion lung scan.

These thromboemboli did not contribute to death. Based on clinical assessment, recurrent PE were thought to have occurred. The frequency of nonfatal recurrent PE was 1 of 20 (5.0%).

Clinical Characteristics of Patients Untreated patients' age was comparable to treated patients (60 ± 18 years vs 58± 17 years) (mean ± SD) (NS). Among untreated patients, 60% were men, and among treated patients, 53% were men (NS). Among 20 untreated patients, pleuritic pain in six (30%) was less frequent than in treated patients (Table 2). All other signs and symptoms occurred with comparable frequency between the two groups. The plain chest radiograph and electrocardiogram also showed no statistically significant differences of the frequency of abnormalities among treated and untreated patients (Table 2) .

Pulmonary Artery Pressure, Pa02, and Alveolar Arterial Oxygen Gradient The PaOz while breathing room air was lower in untreated compared with treated patients (Table 3). The alveolar arterial oxygen gradient while breathing room air was higher in untreated compared with treated patients. The pulmonary artery mean pressure did not show a statistically significant difference between untreated and treated patients (Table 3) .

Ventilation / Perfusion Lung Scans Ventilation/ perfusion lung scans were interpreted as high probability in a smaller percent of untreated patients with PE than treated patients, 0 of 20 (0%) vs 160 of 376 (43%) (p<0.001) . Low probability, nearly normal, or normal ventilation / perfusion scans

Table 3-Pulmonary Artery Mean Pressure, Pa02, and Alveolar Arterial Oxygen Gradient Among Untreated and Treated Patients With Pulmonary Embolism*

Pulmonary artery mean pressure, mm Hg Pa02 , mm Hg A-a gradient, mm Hg

Untreated

Treated

p Value

23±13 (n=19) 55±21 (n=l6) 55±31 (n=l6)

24±10 (n=359) 68±15 (n=272) 39±16 (n=272)

NS <0.01 <0.001

*Pa0 2 =partial pressure of oxygen in arterial blood; A-a gradient=alveolar arterial oxygen gradient. Results are mean± SD.

were more frequent among untreated patients, 10 of 20 (50%) vs 58 of 376 (15%) (p<0.001). Ventilation/ perfusion lung scans among untreated patients more often showed no mismatched segmental equivalent perfusion defects than among treated patients, 14 of 20 (70%) vs 122 of 376 (32%) (p<0.01) (Table 4) . All untreated patients, 20 of 20 (100%), showed fewer than three mismatched segmental equivalent perfusion defects compared with 227 of 376 (60%) among treated patients (p<0.001) .

Pulmonary Angiography Pulmonary angiograms at the time of PIOPED entry were obtained in 19 untreated patients and 362 treated patients. Thromboemboli involved only segmental pulmonary arteries or smaller branches in 16 of 19 (84%) untreated patients and 132 of 362 (36%) treated patients (p<0.001) (Table 4). Only peripheral vessels were involved in 7 of 19 (37%) untreated patients and 13 of 362 (4%) treated patients (p<0.001) . Thromboemboli were not observed on the angiogram of one untreated patient, but were shown at autopsy 6 days later.

Comparison of Results- Untreated and Treated Patients Death from the initial PE or fatal recurrent PE during the first 3 months after the initial PE occurred in 1 of 20 (5.0%) untreated patients and 9 of 376 (2.4%) treated patients (NS) (Table 5). Nonfatal recurrent PE during the 3 months of follow-up occurred in 1 of 20 (5.0%) untreated patients compared with 13 of 376 (3.5%) treated patients (NS) . Total events (recurrent nonfatal PE, fatal recurrent PE, or death from the initial PE) occurred in 2 of 20 (10.0%) untreated patients compared with 22 of 376 (5.9%) treated patients (NS) . There was no statistically significant difference of the frequency of fatal PE, fatal recurrent PE, and nonfatal recurrent PE among untreated patients and patients treated with full-dose anticoagulants, but not thrombolytic therapy , inferior vena cava interruption, or pulmonary embolectomy (Table 5). CHEST / 107 I 41 APRIL, 1995

933

Table 4-Ventilation/Perfusion Lung Scans and Pulmonary Angiograms Among Untreated and Treated Patients Treated

Untreated

No mismatched segmental eq uivalen t perfusion defects (V / Q )* <3 Mismatched segmental equivalent perfusion defects (V / Q)* Segmental pulmonary artery or smaller (angiogram ) Peripheral arteries on! y (angiogram)

No.

(%)

No.

(%)

p Value

14/ 20

(70)

122/ 376

(32)

<0.01

20/ 20

(100)

227/ 376

(60)

<0.001

16/ 19

(84)

132/ 362

(36)

<0.001

7/ 19

(37)

13/ 362

(4)

<0.001

*V / Q=ventilation/ perfusion lung scan.

19 (10.5%) compared with 32 of 376 (8.5%) among patients who received any treatment and 20 of 297 (6.7%) among patients who received full-dose anticoagulants, and perhaps antiplatelet agents in addition , but not thrombolytic therapy, inferior vena cava interruption, or embolectomy. The frequency of events comparing untreated vs treated patients was not statistically significant.

Among patients with no mismatched segmental equivalent perfusion defects shown on the ventilation / perfusion lung scan, fatal initial PE, fatal recurrent PE, or nonfatal recurrent PE during the first 3 months after the initial PE occurred in 1 of 14 (7.1%) untreated patients compared with 6 of 122 (4.9%) treated patients (NS) (Table 5) . Among patients with fewer than three mismatched segmental equivalent perfusion defects, fatal and nonfatal events occurred in 2 of 20 (10.0%) untreated patients compared with 13 of 227 (5.7%) treated patients (NS). Among patients with only segmental arteries or smaller branches showing thromboemboli on the pulmonary angiogram, fatal and nonfatal events occurred in 1 of 16 (6.3%) untreated patients vs 6 of 132 (4.5%) treated patients (NS). In regard to the course of untreated PE over 1 year, 0 of 19 untreated patients suffered recurrent PE during the fourth through 12th months of observation. The frequency of death from the initial PE, fatal recurrent PE, and nonfatal recurrent PE over the course of 1 year among untreated patients was 2 of

DISCUSSION

Pulmonary angiograms and ventilation/ perfusion lung scans showed involvement of smaller vessels and fewer perfusion defects among untreated patients compared with treated patients. The pulmonary artery mean pressure did not differ to a statistically significant extent among untreated and treated patients, but the Pa02 was lower among untreated patients compared with treated patients, and the alveolar arterial oxygen gradient was higher among untreated patients. The course of untreated PE in these patients appears to represent the course of mild PE. The fre-

Table 5-Results of No Treatment vs Treatment Among Patients Stratified According to Severity of PE and All Patients* Treated Patients, No. of Patients With PE / Total No. of Patients (%)

Untreated Patients, No. of Patients With PE/ Total No. of Patients (%) Fatal Initial or Fatal Recurrent PE No mismatched segmental equivalent on V/ Q < 3 Mismatched segmental equivalent on V/ Q ::S Segmental artery on angiogram All patients Patients treated with heparin, oral anticoagulants, or both

Nonfatal Recurrent PE

Fatal or Nonfatal PE Event

Fatal Initial or Fatal Recurrent PE

Nonfatal Recurrent PE

Fatal or Nonfatal PE Event

1/ 14 (7.1)

0/ 14 (O)

1/ 14 (7.1)

4/ 122 (3.3)

2/ 122 (1.6)

6/122 (4.9)

1/ 20 (5.0)

1/ 20 (50)

2/ 20 (100)

7/ 227 (3.1 )

6/227 (2.6)

13/ 227 (5.7)

0/ 16 (O)

1/ 16 (6.3)

1/ 16 (6.3)

2/ 132 (1.5)

4/ 132 (3.0)

6/ 132 (4.5)

1/ 20 (5.0)

1/ 20 (5.0)

2/ 20 (100)

9/ 376 (2.4) 5/ 297 (1.7)

13/ 376 (3.5) 16/ 297 (20)

22/ 376 (5.9) 11 / 297 (3.7)

*All differences are NS. V/ Q=ventilation/ perfusion lung scan.

934

Outcome Assessment of Untreated PE Patients (Stein, Henry, Relyea)

quency of fatal PE in these patients (5.0%) is strikingly lower than the mortality from PE reported in past decades.l· 2 Hermann et al, 2 based on autopsy findings, calculated a 37% mortality from the initial PE. The treatment of these patients was not reported. Barritt and Jordan, 1 among untreated patients, reported a 26% mortality from the initial PE, although some of these patients perhaps died of recurrent PE. Regarding recurrent PE among untreated patients, Hermann and associates 2 reported a 36% frequency of fatal recurrent PE and a 21% frequency of nonfatal recurrent PE. The lower mortality from the initial PE that we observed in untreated patients reflects the less severe PE. The lower frequency of recurrent PE that we observed presumably also relates to the milder severity of PE in these patients. The mortality of untreated mild PE was comparable to the mortality from fatal PE in untreated patients with subtle deep venous thrombosis, about 5%. 6 The 3-month duration of follow-up that we primarily evaluated is the duration recommended for anticoagulant therapy in the absence of continuing risk factors. 7 The frequency of death from PE or nonfatal recurrent PE during this period was comparable among untreated and treated patients stratified according to the severity of the perfusion defect or size of vessels involved on the pulmonary angiogram. Inadequate treatment with antithrombotic agents had a negligible adverse effect on the results of treatment; only 4 of 376 (1.1%) treated patients received antithrombotic agents without the intention

of achieving full anticoagulation. The statistically insignificant difference of fatal initial PE, fatal recurrent PE, or nonfatal recurrent PE among untreated patients with mild PE compared with treated patients with mild PE suggests a possible need for further evaluation of the risk-benefit ratio of anticoagulation treatment of mild PE. ACKNOWLEDGMENT: We thank Jeffrey Carson, MD, and Amy.Duff, PhD, for their assistance in identifying the untreated patients in PIOPED. REFERENCES

1 Barritt DW, Jordan SC. Anticoagulant drugs in the treatment of pulmonary embolism: a controlled trial. Lancet 1960; l: 1309-12 2 Hermann RE, Davis JH, Holden WD. Pulmonary embolism: a clinical and pathologic study with emphasis on the effect of prophylactic therapy with anticoagulants. Am J Surg 1961; 102:19-28 3 Carson JL, Kelley MA, Duff A, et a!. The clinical course of pulmonary embolism. N Engl J Med 1992; 326:1240-45 4 A collaborative study by the PIOPED Investigators. Value of the ventilation/ perfusion scan in acute pulmonary embolism: results of the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED). JAMA 1990; 263:2753-59 5 Stein PD, Terrin ML, Hales CA, et a!. Clinical, laboratory, roentgenographic and electrocardiographic findings in patients with acute pulmonary embolism and no pre-existing cardiac or pulmonary disease. Chest 1991; 100:598-603 6 Collins R, Scrimgeour A, Yusuf S, eta!. Reduction in fatal pulmonary embolism and venous thrombosis by perioperative administration of subcutaneous heparin. N Eng! J Med 1988; 318:1162-73 7 Hyers TM, Hull RD, Weg JG. Antithrombotic therapy for venous thromboembolic disease. Chest 1992; l02(suppl):408S425S

CHEST 1 107 I 4 I APRIL, 1995

935