With written exemption from our institutional review board for this study, we performed a search of our radiology database for pulmonary angiography performed between January 1, 1992, and May 31, 2002. After reviewing the procedural reports, we filtered this query to obtain only those patients with pulmonary hypertension. Pulmonary hypertension was defined as a systolic pulmonary artery pressure (before injection of contrast agent) of greater than or equal to 30 mm Hg (systolic) or 20 mm Hg (mean) [15].

Variables collected for each patient from chart histories and clinical notes specific to the pulmonary angiography that was performed are listed in Appendix 1.

View Larger Version

APPENDIX 1. Variables Collected for Each Patient from Chart Histories and Clinical Notes Specific to the Pulmonary Angiography That Was Performed

The commonly used preoperative American Society of Anesthesiologists (ASA) physical status classification was retrospectively applied to each patient of the study population to categorize them according to their health state at the time of pulmonary angiography [16, 17]. The patients were graded from classes 1 through 5 according to the ASA classification (Appendix 2). This classification scheme is used clinically at our institution before the administration of conscious sedation for invasive procedures. We could not apply a critical illness scoring scheme such as the Acute Physiology and Chronic Health Evaluation (APACHE II) index to these patients because not all laboratory values were available [18]. Although the ASA scoring system is more subjective than the APACHE II index, the ASA system has been shown to correlate well with surgical outcome [19].

View Larger Version

APPENDIX 2. American Society of Anesthesiologists Physical Status Classification: Adapted from Institutional Interventional Radiology Critical Pathway Sheet [ 19 ]

Standard femoral or jugular vein access was established with the placement of a 5- to 7-French vascular sheath. Using standard diagnostic catheters (pigtail catheter, Grollman catheter, Van Aman catheter, Berman angiography catheter) at the discretion of the attending physician, we selectively cannulated the main right or left pulmonary artery. Contrast material injection rates and volumes were not routinely documented but ranged from 10- to 25-mL/sec for a total of 20–50 mL per injection. Additional selective and subselective injections were performed at the discretion of the attending physician. Before contrast injection, pulmonary pressure measurements were obtained. All patients received nonionic contrast material, with more than 98% receiving iohexol (Omnipaque, Amersham Health) and the remainder receiving iodixanol (Visipaque, Amersham Health). When possible, pressure measurements were obtained again after diagnostic angiography.

The pulmonary artery hemodynamic effects of pulmonary angiography were analyzed. We examined the change in pulmonary artery pressure after contrast injection across continuous and noncontinuous data variables. Data analysis was performed using STATA version 7.0 (Statacorp). Continuous variables (age and pressure measurements) were analyzed using Student's t tests and simple linear regressions. Noncontinuous data (race, sex, clinical indication, hypertension severity, hypertension classification, and incidence of complication) were analyzed using the Fisher's exact test. Differences were considered significant if the p value was less than 0.05.

Over a 10-year period, 612 pulmonary angiograms were obtained at our institution. Two hundred two of these patients had pulmonary hypertension as defined previously: 155 had moderate pulmonary hypertension and 47 had severe pulmonary hypertension. A baseline comparison of patient subgroups with moderate and severe pulmonary hypertension is summarized in Table 1. A large number of the patients were suspected of having chronic thromboembolic disease. This is reflective of our referral pattern because we practice at a tertiary care center. The distribution of pulmonary artery pressure measurements obtained before contrast injection in our patient population is depicted in Figure 1. The initial precontrast injection pressure was obtained from the main (n = 67), left (n = 86), or right pulmonary artery (n = 49). The final postcontrast pressure measurement was obtained from the main (n = 20), left (n = 32), or right pulmonary artery (n = 36).

View Larger Version

TABLE 1 Baseline Patient Data by Severity of Pulmonary Arterial Hypertension

View larger version (11K)

Fig. 1. —Bar graph shows distribution of pulmonary artery pressure before injection of contrast agent.

Complications.—As shown in Tables 2 and 3, in our study the overall complication rate was 2.0% (n = 4). All were major complications. Three complications (cardiac arrests) were fatal, and one complication (severe chest pain that required termination of the examination) was nonfatal. The fatal complications occurred in patients with acute pulmonary hypertension, and the nonfatal complication occurred in a patient with chronic pulmonary hypertension.

View Larger Version

TABLE 2 Outcome Variables by Severity of Hypertension

View Larger Version

TABLE 3 Patients with Complications

Patients with severe pulmonary hypertension (systolic pulmonary artery pressure, ≥ 60 mm Hg) had a major complication rate of 6.3% (n = 3) and no minor complications. Conversely, patients with moderate pulmonary hypertension had a complication rate of 0.6% (n = 1) with one major complication (fatal cardiac arrest) and no minor complications.

No significant difference was found in the mean ASA grade between patients with severe pulmonary hypertension and those with moderate pulmonary hypertension (3.2 vs 3.0, respectively). However, the mean ASA score of the patients with complications was higher than that of patients without complications (4.0 vs 3.0, p = 0.03).

Given the low complication rate observed in our study, we thought it might be helpful to describe the clinical scenario for each complication:

Complication 1.—A 53-year-old white woman with 2 years of progressively worsening dyspnea on exertion was admitted to the hospital for recent onset of chest pain. Before the procedure, the patient's ASA score was 3. The precontrast systolic pulmonary artery pressure was 85 mm Hg with no change in pressure after bilateral pulmonary angiography with the administration of 75 mL of contrast material (the radiology report does not give injection rates in the case). The study was negative for pulmonary embolus. After removal of the sheath, the patient complained of severe chest pain similar to the pain experienced on admission. The patient had no ECG changes and no elevation of serial serum cardiac enzymes. The patient was released from the hospital 2 days later without further incident, with a presumptive diagnosis of primary pulmonary artery hypertension. The cause of her two episodes of chest pain was never elucidated.

Complication 2.—A 17-month-old white girl with acute lymphocytic leukemia was initially admitted to the pediatric intensive care unit for tachypnea and presumed left lower lobe pneumonia. A chest CT scan showed a partially occluded left pulmonary artery. The patient became bradycardic and pulseless and was revived by cardiopulmonary resuscitation. In spite of a grave prognosis according to the intensive care physician and an ASA score of 5, we obtained an emergent left pulmonary artery angiogram. The precontrast systolic pulmonary artery pressure was 68 mm Hg; no postcontrast pressure was recorded after a 10-mL contrast hand injection. (Given the age of this patient, this 10-mL volume is probably equivalent to a 50- to 60-mL injection in an adult.) The angiogram showed complete occlusion of the left pulmonary artery. Catheter-directed urokinase (Abbokinase, Abbott Laboratories) was empirically administered. The patient then had two consecutive cardiac arrests involving either complete heart block or pulseless electrical activity. The procedure was aborted; the patient was revived and sent immediately to the pediatric intensive care unit, where she again became bradycardic and pulseless and died. At autopsy, a large thrombus was found occluding the left pulmonary artery. The cause of death was acute right ventricular failure.

Complication 3.—A 32-year-old white woman with preterm labor developed chest pain and dyspnea after a spontaneous vaginal delivery. Echocardiography showed a dilated right atrium, a dilated right ventricle, and increased right ventricular systolic pressure. A Swan-Ganz catheterization measured a cardiac output of 1.6 L/min. Concerned about a pulmonary embolus, we referred the patient for pulmonary angiography. Immediately before the procedure, her ASA score was 4, and the patient experienced supraventricular tachycardia with hypotension that responded to adenosine. During placement of the catheter into the left pulmonary artery, the patient had another episode of hypotension and supraventricular tachycardia, which was converted to sinus tachycardia after administration of multiple doses of adenosine. After the patient was stabilized, the precontrast systolic pulmonary artery pressure was 80 mm Hg. The radiology report stated that only hand injections of the left pulmonary artery were performed with a total of 50 mL of contrast material. The angiogram was negative for pulmonary embolus. No postcontrast pressure was recorded. During attempts to access the right pulmonary artery, the patient became agitated, apneic, bradycardic, and hypotensive. The patient then went into a junctional rhythm and died, in spite of resuscitative efforts. Postmortem findings were idiopathic primary pulmonary arterial hypertension and no evidence of pulmonary embolus. The cause of death was acute right ventricular failure.

Complication 4.—A 45-year-old black woman with metastatic breast cancer presented to the hospital with a few days of worsening dyspnea with concomitant hypoxia (oxygen saturation, 81% by pulse oximetry on room air). After undergoing ventilation–perfusion nuclear scanning showing intermediate-probability results, the patient was started on heparin and referred for pulmonary angiography. The patient's preprocedural ASA score was 4. The precontrast systolic pulmonary artery pressure was 58 mm Hg. Left pulmonary artery angiograms in the anteroposterior (20 mL/sec for a volume of 40 mL) and oblique projections (20 mL/sec for a volume of 40 mL) were successfully obtained; there was no evidence of a pulmonary embolus. No pressure measurement was recorded after contrast administration. Before the right side could be selected, the patient vomited and precipitously became more hypoxic and apneic. After an oral airway was placed, the patient seized, went into pulseless electrical activity, and then became asystolic. In spite of the administration of cardiopulmonary resuscitation and use of an emergent transvenous pacer, the patient remained pulseless and died. The autopsy report indicated massive intramicrovascular pulmonary metastases. The cause of death was acute right ventricular failure.

Because the incidence of complications can be used as an objective measure of the safety of a procedure, we decided to examine whether any of the data variables have a predictive value in determining the occurrence of complications during pulmonary angiography. To do this, we separated the patients into two groups: those with and without complications (Table 4). We then analyzed our outcome variables for differences across these two groups (Table 5).

View Larger Version

TABLE 4 Baseline Patient Data by Complication

View Larger Version

TABLE 5 Outcome Variables by Complication

The pulmonary artery hemodynamic changes are listed in Table 6. We found that as the baseline preangiography pulmonary artery pressure increases, the observed change in pulmonary artery pressure decreases; quantitatively, the observed change in pressure becomes approximately 1 mm Hg less per 10-mm-Hg increase in initial preinjection pulmonary artery pressure. Patients whose clinical indication was previous lung transplantation showed significantly higher increases in pulmonary artery pressure relative to all other indications combined (16.75 ± 12.97 mm Hg vs 5.46 ± 6.86 mm Hg, p = 0.003). The remaining data variables were not statistically significant for change in pulmonary artery systolic pressure after angiography.

View Larger Version

TABLE 6 Pulmonary Artery Pressure Changes Based on Baseline Data

The continuous data variables were analyzed for significance with regard to change in pulmonary artery pressure after contrast injection. No significant correlations were found between the change in pulmonary artery pressure after angiography and age (p = 0.2), contrast volume (p = 1.0), or pulmonary artery pressure before angiography (p = 0.1) or after angiography (p = 0.08).

In our series of patients with pulmonary hypertension, the overall number of complications was low but accounted for 2% of the study population. Prior reports document an overall mortality rate of pulmonary angiography of approximately 0.3% in patients with or without pulmonary artery hypertension, with most of these deaths occurring from heart failure, cardiac arrhythmia, or severe allergic reaction (including renal failure) to contrast media [20]. A prior study of patients with pulmonary hypertension, using ionic and nonionic contrast material, reported no fatalities in 67 patients [14, 21]. However, the mortality rate in our study was 1.5%.

Patients with complications were more seriously ill with higher mean ASA scores than those without complications. All fatal complications were due to acute right ventricular failure and occurred in patients with acute pulmonary hypertension. Patient 2 had a massive thrombus occluding the left pulmonary artery, and patient 4 had massive intramicrovascular metastases. Patient 3 had primary pulmonary hypertension and was pregnant; pregnancy exacerbates primary pulmonary hypertension and is associated with a greater than 50% mortality rate due to estrogen-mediated vasoreactive changes [22]. In all these patients, significant right ventricular strain was present before angiography because of increased right ventricular afterload. After contrast administration, the right ventricular afterload increased to a point at which right ventricular failure ensued.

When stratified into severe and moderate pulmonary hypertension, no difference in mean ASA scores was found between the two groups, but a higher incidence (6.3%) of major complications was seen in patients with severe pulmonary hypertension (systolic pulmonary artery pressure, ≥ 60 mm Hg) compared to a 0.6% incidence in patients with moderate pulmonary hypertension (p = 0.63). Although this difference was not statistically significant, further analysis revealed an interesting pattern. Intuitively, we would expect patients with the highest pulmonary artery pressures to have the most complications. However, in our series, none of the nine patients with pulmonary artery pressures greater than 90 mm Hg had complications. The four complications occurred in patients with systolic pressures between 58 and 85 mm Hg. Thus, the trend was for more complications in patients with higher systolic pulmonary artery pressures but not necessarily in the group of patients with the very highest pressures.

Previous studies have shown pulmonary artery hemodynamic changes after pulmonary angiography [4–7]. In particular, contrast administration has been shown to be associated with an increase in pulmonary artery pressure [16, 23]. The pulmonary artery pressure effects are related to cardiac output and pulmonary vascular resistance (PVR) through the formula [24]:

The increase in the pulmonary artery pressure observed after pulmonary angiography can potentially be due to either an increase in the PVR; an increase in cardiac output, such as from the effects of the contrast agent in lowering systemic vascular resistance; or an increase in left ventricular filling pressure, which could be due to the increased fluid load associated with contrast administration.

Mixed opinions are found in the literature regarding the effects of contrast administration on peripheral vascular resistance and cardiac output. Some investigators have observed that contrast material increases pulmonary artery pressure via a large increase in cardiac output resulting from decreased systemic vascular resistance, in spite of a concomitant decrease in PVR [25, 26]. Others, however, have observed that the increased pulmonary artery pressure was due to a sustained increase in PVR after an immediate initial decrease in PVR [27]. However, all these studies were performed in patients or animals with normal pulmonary artery pressures.

When examining patients with pulmonary hypertension, researchers have proposed that such patients have a disease-induced impairment of pulmonary vasodilatation capacity [28]. This deficiency is exacerbated by contrast agents that not only increase the volume load but also increase PVR through vasoconstriction [28]. We found a greater increase in pulmonary artery pressure in patients with moderate pulmonary hypertension than in those with severe pulmonary hypertension after contrast injection. We hypothesized that part of the increase in pulmonary artery pressure may be due to pulmonary vasoconstriction and that those patients with severe pulmonary hypertension may have already reached near maximal vasoconstriction, whereas those patients with moderate pulmonary hypertension had more of a dynamic range related to vasoconstriction. Moreover, we found that the rise in pulmonary artery pressure after contrast injection was greater in patients after lung transplantation than in all other clinical indication groups combined. Again this rise may be due to contrast-induced vasoconstriction. The baseline pulmonary artery pressure may be elevated because of the high vascular tone in the nontransplanted lung, but the increase in pressure seen after contrast injection may be due to contrast-induced vasoconstriction in the “good” transplanted lung.

It has also been reported that there is a greater increase in cardiac output after contrast administration in patients with moderate pulmonary hypertension compared with patients with severe pulmonary hypertension [28]. This may possibly explain the greater increase in pulmonary artery pressure among patients with moderate pulmonary hypertension in our series. Because we did not attempt to measure PVR or cardiac output during our pulmonary angiography, we could not evaluate differences in the effect of the contrast administration on PVR and cardiac output between these two groups of patients. Similarly, we could not determine the cause of the difference in pulmonary artery pressure change that we observed between the two groups.

In our series of patients with pulmonary artery hypertension who underwent pulmonary angiography, we observed a 2% complication rate. All fatal complications were in patients with acute pulmonary hypertension, and patients with complications had a higher mean ASA score than those without complications. In the setting of chronic pulmonary hypertension, pulmonary angiography was relatively safe. On the basis of these findings, we recommend that severely ill patients with elevated pulmonary artery pressures, particularly acute pulmonary hypertension, be treated with particular caution; in this subset of patients, every noninvasive imaging method should be exhausted before proceeding to pulmonary angiography.