点击显示 收起
ABSTRACT |
---|
TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES |
---|
Key Words: pulmonary hypertension • mortality • flolan
Pulmonary arterial hypertension (PAH), a disease primarily affecting the small precapillary pulmonary vessels, is characterized by sustained elevation of the pulmonary vascular resistance (PVR). Without therapy, right heart failure and death eventually occur (1–3). PAH occurs in an idiopathic form, primary pulmonary hypertension (PPH), and in association with other disorders such as connective tissue diseases or congenital heart disease (CHD). Although a genetic defect in bone morphogenetic protein receptor 2 has been found in many patients with sporadic and familial PPH, PAH remains a disease of unknown etiology (4). Until recently, prognosis for PAH was poor, with a mean survival of less than 3 years (1).
Prostacyclin (prostaglandin I2), produced primarily by the vascular endothelium, is a potent vasodilator and inhibitor of platelet aggregation and smooth muscle growth (5, 6). Beneficial long-term treatment with the synthetic salt of prostacyclin, epoprostenol, was first reported in 1983 in a patient with PPH (7). Epoprostenol is given by continuous intravenous infusion due to its half-life of 3 to 5 minutes in vivo and is a difficult and potentially dangerous therapy. Nonetheless, epoprostenol was shown to improve exercise capacity, pulmonary hemodynamics, and 12-week survival in PPH in a randomized, multicenter, trial in 1996 (8). Epoprostenol also improves hemodynamics, exercise capacity, and quality of life in PAH associated with CHD and connective tissue disease (9–11). Survival benefit, however, has only been demonstrated for PPH.
Until recently, the survival benefit from epoprostenol had not been well characterized. A number of studies in the 1990s evaluating the effects of epoprostenol in patients with PPH reported improved outcome but were limited by small sample size (12, 13), lack of follow-up hemodynamics (7), and evaluation of only echocardiographic data (14). Two recent publications by McLaughlin and associates and Sitbon and associates, have convincingly demonstrated long-term survival benefits in two large cohorts of patients with PPH (15, 16). No study has reported the effect of epoprostenol on long-term survival in other forms of PAH. The purpose of this study was to determine the long-term effect of epoprostenol on mortality and to determine factors associated with outcome in patients with different etiologies of PAH.
METHODS |
---|
TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES |
---|
Epoprostenol was initiated after either failure to improve clinically with calcium channel blocker therapy or failure to demonstrate a significant decrease (> 20%) in mean pulmonary arterial pressure at the time of acute vasodilator testing. Epoprostenol was administered through a permanent venous catheter using a CADD-1 portable infusion pump (Model 5100HF; Pharmacia Deltec Inc., St. Paul, MN). Patients were admitted to the hospital for an average of 3 to 5 days to initiate therapy. Average dose of epoprostenol at hospital discharge was 4 to 6 ng/kg/minute based on patient admission weight. After discharge, epoprostenol dose was initially increased weekly by 1 ng/kg/minute as limited by side effects (headache, nausea, diarrhea, jaw pain, and foot pain) with a goal of 20 ng/kg/minute at 4 to 6 months. The epoprostenol administration regimen did not change significantly during the study period.
Hemodynamic data was obtained using a 7.0 or 7.5 French pulmonary artery thermodilution catheter (Baxter, Irvine, CA) or 7.0 French balloon-tipped end-hole catheter (Arrow International, Reading, PA). Cardiac output was calculated using the Fick equation, thermodilution technique, or both. PVR was calculated using the standard formula (20). Hemodynamic data was routinely obtained at baseline and 1 year (± 6 months) after starting epoprostenol.
Statistical Analysis
All continuous measures are reported as mean ± SD, and categoric measures are reported as proportions. Age of disease onset was defined as the onset of symptoms as noted in the initial medical records. Differences in patient characteristics between etiology groups were evaluated using the 2 test for categoric measures and analysis of variance for continuous measures. Within each etiology group, differences in patient hemodynamics at baseline versus 1 year after epoprostenol treatment were evaluated using the paired t test. Nonparametric tests were used as appropriate. Kaplan-Meier plots were used to estimate survival distributions. The log rank test was used to compare survival curves between etiology and other patient characteristics. Cox proportional hazards regression was used to determine baseline covariates related to survival. Length of survival was censored at death, transplantation, or time of last contact. Regression diagnostics were used to assess the proportional hazards assumption. Proportional hazards regression with time-dependant covariates was used to determine predictors of survival measured at baseline and again at follow-up. A two-sided p value of less than 0.05 was considered statistically significant. Statistical analyses were performed using SPSS (release 10.0.7; Chicago, IL) and SAS (release 8.1; Cary, NC). Further details are available in the online supplement.
RESULTS |
---|
TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES |
---|
|
|
Epoprostenol Administration
Mean epoprostenol dose at 1 year was 23 ± 18 ng/kg/minute, median dose was 18 ng/kg/minute. Mean follow-up time was 2.4 ± 1.8 years after initiating epoprostenol. Local Hickman catheter site infections that did not respond to antibiotics and required replacement of the Hickman catheter occurred in 10% of our patients. Two patients developed sepsis related to catheter infections, and one patient developed septic emboli from a catheter infection. All three patients had resolution of the infection with antibiotics and replacement of the catheter. Two CHD patients with unrepaired defects initially had double-lumen Hickman catheters placed and experienced paradoxical air-emboli (without neurologic changes) when flushing the unused port. These events resolved after the catheter was changed to a single-lumen catheter. No patient died of complications related to the delivery system for epoprostenol.
Functional Status
Before starting epoprostenol, 47 patients were functional Class III and 44 patients were functional Class IV. After 1 year of treatment, 19 patients had died, of whom 5 patients were initially functional Class III and 14 were initially functional Class IV, and 2 others received lung transplant. Of the remaining 70 patients, 61 (88%) were functional Class I, II, or III. There were no significant differences in baseline functional class by etiology. World Health Organization (WHO) functional class and mortality 1 year after initiating epoprostenol is presented in .
|
Hemodynamics
Baseline and follow-up hemodynamics are shown in . Twenty-nine patients did not undergo repeat hemodynamic evaluation within 12 ± 6 months. Of these patients, 19 died less than 1 year after initiating epoprostenol, 8 patients underwent repeat hemodynamic evaluation more than 18 months after initiation of therapy (4 were started on epoprostenol at outside institutions where repeat hemodynamic evaluation was not routine, 4 declined hemodynamic evaluation, 1 patient received transplantation 3 months after starting epoprostenol, and 1 patient had no repeat hemodynamic evaluation due to comorbid illness).
|
Survival
The proportion of patients surviving 1, 2, and 3 years was 79, 70, and 59%, respectively. The 1-, 2-, and 3-year survival in patients with PPH was 85% (95% confidence interval [CI] 72–93%), 76% (95% CI 61–86%) and 65% (95% CI 47–78%), respectively. Their 1-, 2-, and 3-year survivals predicted by the National Institutes of Health PPH survival equation were 62, 49, and 39%, respectively. Conversely, survival in patients with SSD was 58% (95% CI 33–76%), 41% (95% CI 19–61%), and 34% (95% CI 14–55%) at 1-, 2-, and 3-years, respectively. Survival of patients with PPH compared with other etiologies is shown in . Of the 37 deaths that occurred during the study period, 29 were due to worsening right heart failure, 6 were due to sudden death, and 2 were due to a massive gastrointestinal bleed (both patients had SSD, neither was receiving chronic anticoagulation).
|
|
|
DISCUSSION |
---|
TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES |
---|
Long-term outcome in patients with SSD has not been reported in a large study. In the only study looking exclusively at patients with SSD, Klings and associates reported only three deaths in 12 patients receiving long-term epoprostenol therapy followed for up to 27 months (21). McLaughlin and colleagues reported only two deaths in 14 patients with connective tissue disease–associated pulmonary hypertension after 12.7 ± 5.6 months of epoprostenol treatment with a significant hemodynamic improvement (mean decrease in pulmonary arterial pressure of 23% and in PVR of 46%) (22). Conversely, in a study of 17 patients with connective tissue disease, 9 of whom had SSD, Humbert and coworkers reported a 53% mortality at 80 ± 48 weeks after starting epoprostenol (23).
Despite hemodynamic improvement similar to that seen in patients with PPH, our patients with SSD had a much worse survival. This is in agreement with the absence of survival benefit noted in an earlier randomized, multicenter study comparing short-term (3-month) epoprostenol therapy with conventional therapy in patients with SSD (10). Although our study represents the largest long-term follow up of patients with SSD treated with epoprostenol, there is a need for additional studies with greater numbers of patients.
Why patients with SSD had a worse outcome likely involves many factors. They were significantly different from the rest of the cohort in several ways: they were older (an average of 15.3 years older compared with PPH patients, at the time of disease onset), used more calcium channel blockers, and had a lower diffusing capacity of carbon monoxide. However, the mean age of disease onset of our patients with SSD was similar to that in previous studies, and even after adjusting for age, the diagnosis of SSD remained an independent risk factor for worse outcome (21, 23). Although occult interstitial lung disease cannot be excluded as a contributing factor, this appears to be unlikely. Total lung capacity was not reduced compared with other etiologies, and epoprostenol was not offered to patients with evidence of more than mild interstitial changes on high-resolution computed tomography of the chest. In addition, even after adjusting for decreased diffusing capacity of carbon monoxide, the diagnosis of SSD was associated with a poorer outcome.
One possibility is the systemic nature of SSD. Epoprostenol may improve the pulmonary hemodynamics in SSD while not altering the ongoing systemic illness. However, worsening right heart failure or sudden death was the cause of death in all but 2 of the 15 SSD patients who died during the course of the study. Another possibility is delayed diagnosis in these patients. Their fatigue and shortness of breath may have initially been attributed to scleroderma rather than pulmonary hypertension. A third possibility is that different disease processes may cause PPH and SSD-associated pulmonary hypertension. Although the pathologic lesions are similar in PPH and other forms of PAH, Lee and associates demonstrated monoclonal endothelial cell proliferation in plexiform lesions of patients with PPH, whereas there was polyclonal cell proliferation in patients with SSD and CHD (24). These changes may account for the different outcomes achieved. Finally, the possibility of "pulmonary Raynaud's"—additional vasospasm over and above fixed disease—contributing to pulmonary hypertension, seems unlikely. None of the 19 patients with SSD exhibited an acute vasodilator response at any time, and we have yet to see a patient at our institution with SSD-associated PAH improve with calcium channel blocker therapy.
The specific type of connective tissue disease may be important because all patients in this cohort with SLE are still alive with sustained clinical improvement. As noted in an earlier case series, epoprostenol was well tolerated in patients with SLE (25). However, this has not been the experience of all investigators. Horn and associates have reported significant side effects, in particular severe thrombocytopenia, with minimal benefit (26).
Survival and hemodynamic improvement in the 11 patients with CHD was similar to that seen with PPH, despite the fact that CHD patients had documented pulmonary hypertension for a much greater period of time. This implies that even with long-standing pulmonary hypertension, therapeutic benefit may be possible.
Epoprostenol therapy was well tolerated by patients with CHD, but it is unclear if epoprostenol affects their long-term survival. There are few published reports on outcome with epoprostenol therapy in CHD-related pulmonary hypertension. McLaughlin and associates reported significant hemodynamic improvement in seven patients with CHD with only one death, although five patients had previously undergone surgical correction of their defect (22). Rosenzweig and associates published the long-term results of epoprostenol use in 20 patients with CHD-associated PAH, none of who exhibited acute vasoreactivity (9). Nine patients had undergone previous surgical correction of their defect, and only five patients demonstrated a resting peripheral saturation of less than 90%. They found significant improvement in hemodynamics and exercise capacity with an average decrease in of 21%. Most of the patients in our study had significant hypoxemia, indicating greater right-to-left shunting and worse right ventricular function. This may account for the less dramatic hemodynamic improvement that we observed.
With regard to predictors of outcome, in addition to the diagnosis of SSD, two other factors were associated with outcome. Age at disease onset was the most powerful factor. Patients older than the median age of 44 were at significantly increased risk of dying. It is not entirely clear why age should portend a worse outcome. Although coexisting medical problems are more likely to be present in older patients, nearly all the deaths in our series were related to worsening right heart failure or sudden death. Older patients are more likely to have concomitant coronary artery disease and left-sided heart disease, which may be adversely affected by epoprostenol. In the Flolan International Randomized Survival Trial trial, increased mortality was found with epoprostenol use in patients with ischemic heart disease (27). Overdosing of epoprostenol and the subsequent development of high output failure has also been reported (28).
The third factor associated with worse outcome was WHO functional Class IV at any time. This is comparable with the findings of McLaughlin and associates in their survival analysis of 162 PPH patients (15). They found that at 1 year patients improving to Class I or II had an 89% 3-year survival as opposed to those who remained functional Class IV (0% 3-year survival). Similarly, Sitbon and colleagues, evaluating 178 PPH patients over an 8-year period, found that persistence of functional Class III or IV at 3 months was associated with poor survival.
In contrast to these two studies looking only at patients with PPH, we found no hemodynamic factors predictive of outcome in multivariate analysis. McLaughlin and colleagues found that improvement in cardiac index and correlated with improved survival, whereas Sitbon and associates found that patients with less than a 30% decrease in PVR, after 3 months of epoprostenol treatment had a poorer survival (15, 16). They found no improvement in survival associated with cardiac index or .
There are two likely explanations for the lack of hemodynamic predictors of outcome in our study. We analyzed a smaller cohort, and with additional patients, hemodynamic predictors of outcome may have been significant. Indeed, there was a trend toward significance with regard to worse outcome in patients with a decrease in PVR of less than 30% (HR 1.66 [0.57–4.84]). In addition to a smaller sample size, the inclusion of patients whose hemodynamics parameters, at baseline and follow-up, differed from PPH patients, may in part account for our results.
There are a number of potential weaknesses of this study. It is retrospective, and as such, may have biases that we have not taken into account. Although there were no major changes in the clinical management strategy of patients over the 6-year period of the analysis, there may have been subtle changes in care that could have affected outcome. However, the fact that we did not see improvement in survival in SSD, suggests that the improvement in PPH is due to the use of epoprostenol, rather than to better general management of pulmonary hypertension over time. There was also great variety in the length of follow-up of individual patients, but this was accounted for in our statistical analysis. Finally, as in all observational studies, not all records are complete. We were unable to find documentation of baseline pulmonary arterial saturation in seven patients and right atrial pressure in three patients.
In conclusion, epoprostenol prolonged survival in PPH compared with that predicted by the National Institutes of Health registry equation. This was significantly better than the survival seen with SSD-associated PAH, despite the fact that similar hemodynamic improvement was seen in these two groups. Patients with SLE and CHD-associated PAH appeared to benefit from epoprostenol although the numbers in this study are too small to draw firm conclusions. Age at disease onset and functional class were the strongest independent predictors of survival. These will need to be validated in a larger cohort of patients.
Acknowledgments |
---|
REFERENCES |
---|
TOP ABSTRACT METHODS RESULTS DISCUSSION REFERENCES |
---|