Predictors of Failure of Medical Management in Uncomplicated Type B Aortic Dissection

Abstract

Background.

Optimal medical therapy (OMT) for uncomplicated type B aortic dissection (uTBAD) provides excellent short-term outcomes but is associated with a high incidence of failure. This study identified predictors of aortic intervention and mortality in uTBAD patients undergoing OMT.

Methods.

A retrospective review of the Emory University School of Medicine aortic database identified 314 uTBAD patients undergoing OMT from 2000 to 2016. Two hundred sixty-three (84%) patients had imaging at presentation analyzed for maximum aortic diameters (ADs), false lumen (FL) status, and visceral vessel perfusion. Cox proportional hazards models were constructed to estimate hazards ratios (HRs) and identify predictors of OMT failure.

Results.

The mean age of patients was 58 ± 12 years, and 67% were men. FL status was patent in 59.4%, partially thrombosed in 39.8%, and completely thrombosed in 0.8% of patients. Over a median follow-up of 5.6 (interquartile range, 1.4 to 8.5) years, 44.9% of patients failed OMT and underwent intervention (n = 58 open, n = 83 endovascular). The estimated incidence of OMT failure was 46%. Multivariate analysis identified the presence of diabetes, renal failure, DeBakey 3B dissection, and a descending thoracic AD of 4.5 cm or greater (HR, 1.39; 95% confidence interval, 1.24 to 1.56; p < 0.001) to be independent predictors of failure of OMT. FL status or the distribution of visceral vessels arising from the FL did not predict OMT failure.

Conclusions.

There is a significant incidence of OMT failure in uTBAD patients. A descending thoracic AD of 4.5 cm or greater at the time of diagnosis is an independent predictor of failure of OMT.

Optimal medical therapy (OMT) has traditionally been the treatment of choice for patients presenting with acute uncomplicated type B aortic dissection (uTBAD). Currently, surgical or endovascular intervention is reserved for the treatment of complications that arise in the chronic phase. Although OMT is associated with excellent in-hospital and 1-year outcomes, the long-term survival with OMT alone remains poor. Mortality approaches 40% at 5 years, and the intervention-free survival rate is 50% at 5 years and 30% at 10 years [1–3]. Therefore, alternative therapies are necessary to reduce mortality in the chronic phase of TBAD. In a previous report, we demonstrated that thoracic endovascular aortic repair (TEVAR) at the index hospitalization may confer a long-term survival advantage and serve as optimal therapy for both complicated and uncomplicated acute TBADs [4]. Endovascular therapy eliminates antegrade perfusion of the false lumen (FL) and remodels the dissected aorta by expanding the true lumen and shrinking or obliterating the FL [5]. Aortic remodeling with TEVAR has been shown to reduce all-cause and aortic related mortality compared with OMT [3]. However, TEVAR is not a risk-free procedure, and carries periprocedural risks of retrograde type A dissection, stroke, spinal cord ischemia, and acute kidney injury [6].

Therefore, TEVAR should only be utilized in those patients who will ultimately fail OMT in the chronic phase. Currently, the lack of understanding regarding the pathophysiology of TBAD has limited the ability to predict which patients will undergo aortic expansion requiring intervention, and which will be adequately treated with OMT alone. The ability to select those patients who would benefit from early endovascular therapy could dramatically impact long-term survival. The purpose of this study was to analyze predictors of failure of OMT in patients presenting with acute uTBAD over the past 16 years at our institution.

Patients and Methods

This study was conducted with the approval of the institutional review board at Emory University School of Medicine in compliance with HIPAA regulations and the Declaration of Helsinki. The institutional review board waived the need for individual patient consent. The Emory Data Warehouse was queried using the International Classification of Diseases Ninth and Tenth Revision codes for type B aortic dissection performed between 2000 and 2016. After a detailed retrospective review of the electronic medical record, 314 patients with uTBADs who received OMT at index hospitalization were identified. Patients with complicated TBAD, residual distal dissections after proximal aortic repair for type A aortic dissection, and all other acute aortic syndromes involving the descending thoracic aorta were excluded from this analysis. The electronic medical record was reviewed to collect demographics, the date of the initial diagnosis of uTBAD, comorbidities, procedure-specific details, and complications.

The primary endpoint of the study was failure of OMT, which was defined as a composite of aortic intervention or estimated aortic-specific mortality. The indications for surgical or endovascular intervention were aortic growth with a maximum aneurysm size of 5.5 cm, rapid aortic growth (≥5 mm/year), rupture, malperfusion, or intractable pain despite adequate blood pressure control. The decision to use an open or endovascular approach for intervention was surgeon-dependent and based on the indication for intervention, timing from initial presentation, and anatomic factors. The primary outcomes measured were mortality and need for any intervention in the chronic phase.

Imaging Analysis

Patients underwent contrast-enhanced computed tomographic (CT) or magnetic resonance angiogram (MRA) at presentation. Two hundred sixty-three (84%) patients had 1 or more imaging scan at presentation analyzed for maximum aortic diameters, FL status, and visceral vessel perfusion. Aortic centerline analysis was performed using an Aquarius iNtuition 3-dimensional workstation (TeraRecon Inc, San Mateo, CA), and descending thoracic aortic diameters were measured as the maximum diameters orthogonal to the aortic centerline. A patent FL was defined as the absence of thrombus in the FL, complete thrombosis was defined as the absence of contrast in the FL, and partial FL thrombosis was defined as the presence of both thrombus and contrast in the FL. The incidence of completely thrombosed FL in our data was trivial (<1%) and thus was not included in our final analysis. Perfusion status of visceral vessels (celiac artery, superior mesenteric artery, right and left renal arteries, and inferior mesenteric artery) was classified as originating from the TL, the FL, or both. Only cases when visceral vessel perfusion was completely from the FL were counted as “arising from the FL” in our final analysis.

Follow-Up

Follow-up data were obtained via office visits, telephone calls, queries of the Social Security Death Index, or Internet obituary searches. Follow-up within the last 12 months was complete in 86% (270 of 314) of all patients. Every effort was made to determine cause of death for all mortalities in the follow-up period. Nevertheless, although there were 35 confirmed mortalities in the definitive OMT group, cause of death was unknown in 27 of these patients. An estimated rate of aortic-specific mortality was determined after excluding patients with unknown or nonaortic cause of death, and this was included in the final estimation of OMT failure.

Statistics

Categorical variables were summarized using frequencies and percentages. Continuous variables were analyzed with means and standard deviations (SDs) if normally distributed; otherwise medians with interquartile ranges (IQRs) were used. Comparisons between groups were performed using χ² analysis for qualitative variables and Student t test for continuous variables. The primary outcome was time to either mortality or surgical intervention. Because of missing values, multiple imputation was performed, which resulted in 25 complete data sets. Cox proportional hazard regression analysis was then conducted on each data set, which identified the risk factors for death or surgical intervention. The results were consolidated using Rubin’s rules, providing one set of parameter estimates, confidence intervals, and p values. A complete case analysis was also performed, and the conclusions did not differ from those obtained using multiple imputation. When the frequency of any nominal variable was less than or equal to 5, a Fischer exact test was used. Relationships between demographic, clinical, and imaging variables and failure of medical management were assessed using univariate Cox proportional hazards analysis. Variables with statistical significance (p < 0.05) were included in a multivariate model. Survival and intervention rates were estimated using Kaplan-Meier methodology and compared between groups using the log-rank test.

Results

Table 1 lists the baseline demographics and radiographic features of patients with uTBAD at the time of initial presentation. The mean age of all patients was 58 ± 12 years, and 67% were men. There was a high incidence of hypertension (>90%) among all patients, and the majority of patients presented with acute Debakey IIIB dissections. Of the 254 (81%) patients presenting with uTBAD with a contrast scan available for analysis, 151 (59.4%) patients had a patent FL, 101 (39.8%) patients had partial FL thrombosis, and 2 (0.8%) patients had complete FL thrombosis. In 43.1% of patients, there was a single visceral vessel that originated from the FL, and in 30% of patients, multiple visceral vessels originated from the FL. The left renal artery was the most common vessel to originate from the FL. The median follow-up was 5.6 (IQR, 1.4 to 8.5) years.

Table 1.

Baseline Characteristics of Patients Presenting With uTBADs (n = 314)

Variable	All
Age, years	57.9 ± 11.9
Male	67.2 (211 / 314)
Hypertension	93.6 (294 / 314)
Diabetes mellitus	16.2 (51 / 314)
End-stage renal disease	10.2 (32 / 314)
History of stroke	9.9 (31 / 314)
COPD	10.8 (34 / 314)
Active or recent tobacco use	24.5 (77 / 314)
Dyslipidemia	27.7 (87 / 314)
Congestive heart failure	10.2 (32 / 314)
Acute DeBakey III aortic dissection
IIIB	72.3 (227 / 314)
Baseline maximum aortic diameter, cm
Descending aorta	4.6 ± 1.1
Abdominal aorta	3.7 ± 0.9
FL status at time of presentationa
Patent	59.9 (151 / 252)
Partial	40.1 (101 / 252)
Visceral vessel arising from FL
Celiac artery	15.9 (37 / 232)
SMA	6.4 (15 / 233)
Left renal	30.8 (69 / 224)
Right renal	25.0 (56 / 224)
IMA	27.3 (51 / 187)
1 vessel arising from FL	43.2 (102 / 236)
2 vessels arising from FL	14.8 (35 / 236)
3 vessels arising from FL	6.8 (16 / 236)

^aExcluding those with thrombosed (n = 2) or unknown (n = 60) false lumen (FL) status on initial presenting scan.

Values are mean ± SD or percentage (n/N).

COPD = chronic obstructive pulmonary disease; IMA = inferior mesenteric artery; SMA = superior mesenteric artery; uTBAD = uncomplicated type B aortic dissection.

The in-hospital and overall mortality rates for the entire cohort (n = 314) of uTBAD patients were 5.1% and 26.8%. Failure of OMT occurred in 146 of 314 (46%) patients. In the chronic phase, 141 (44.9%) patients exhibited significant aortic expansion requiring either open aortic replacement (n = 58) or endovascular therapy (n = 83). In the remaining 173 patients who received definitive OMT, there were 35 total mortalities during the study period, of which 4 were aortic-related, 4 were non-aortic related, and 27 were due to an unknown cause. The maximum diameter of the descending thoracic aortic (DTA) at presentation was significantly larger (p < 0.001) in patients who failed OMT (5.0 ± 1.0 cm) compared with those patients who remained alive with definitive OMT (4.2 ± 0.9 cm). For patients who failed OMT and required intervention, the maximum DTA diameter increased from 4.8 ± 1.0 cm at baseline to 5.7 ± 1.2 cm prior to intervention.

Predictors of Failure of Optimal Medical Therapy

On univariate analysis, the presence of diabetes mellitus, end-stage renal disease (ESRD), active/recent tobacco use, acute Debakey IIIB versus IIIA dissection, a maximum DTA diameter greater than or equal to 4.5 cm, and partial thrombosis of the FL (versus patent FL) were identified as risk factors for failure of OMT. On multivariate analysis, only diabetes mellitus, ESRD, acute DeBakey IIIB dissection, and aortic size were determined to be independent predictors of failure of OMT (Table 2). In the 254 patients with contrast imaging available for analysis, 136 (54%) patients had DTA diameters greater than or equal to 4.5 cm. In-hospital mortality for these patients was not impacted by aortic diameter. Nevertheless, the incidence of aortic intervention (≥4.5 cm 62.0% versus <4.5 cm 23.8%, p < 0.001) and overall mortality (≥4.5 cm 35.8% versus <4.5 cm, p = 0.003) were significantly higher in patients with DTA diameters greater than or equal to 4.5cm at the time of diagnosis. The overall incidence of failure of OMT in patients with aortic diameters greater than or equal to 4.5 cm was 72.3% (Table 3). Kaplan-Meier estimates of the intervention-free survival in uTBAD patients who presented with DTA diameters greater than or equal to 4.5 cm at 3, 5, and 10 years was 50.1%, 39.3%, and 19.1%, respectively. This was significantly lower than those uTBAD patients who presented with an aortic size less than 4.5 cm and had intervention-free survival rates at 3, 5, and 10 years of 70.6%, 63.4%, and 49.1%, respectively (p < 0.05, Fig 1).

Table 2.

Predictors of Failure of Medical Management on Univariate and Multivariate Analysis (n = 314)

	Univariate Analysis		Multivariate Analysis
Variable	HR (95% CI)	p	HR (95% CI)	p
Age ≧65 years	1.00 (0.71,1.39)	0.990
Male	1.17 (0.85, 1.62)	0.340
Hypertension	0.71 (0.40, 1.25)	0.230
Diabetes mellitus	1.55 (1.34, 1.87)	0.010	1.56 (1.35, 1.89)	0.010
End-stage renal disease	1.50 (1.27, 1.89)	0.020	1.54 (1.30, 1.99)	0.040
History of stroke	1.25 (0.79, 1.97)	0.340
COPD	1.40 (0.91, 2.15)	0.130
Active or recent tobacco use	1.56 (1.13, 2.17)	0.010
Dyslipidemia	0.90 (0.64, 1.25)	0.520
Congestive heart failure	0.93 (0.56, 1.53)	0.770
Acute DeBakey III aortic dissection
IIIB (versus IIIA)	1.59 (1.43, 1.82)	0.002	1.61 (1.44, 1.84)	0.002
Aortic size
Maximum descending aortic diameter ≧4.5 cm	1.40 (1.23, 1.59)	<0.010	1.39 (1.24, 1.56)	<0.001
Maximum abdominal aortic diameter ≧3.5 cm	1.21 (0.99, 1.47)	0.060
FL statusa
Partial	1.43 (1.01, 2.03)	0.045
Visceral vessel arising from FL
Celiac artery	1.00 (0.57, 1.74)	0.996
SMA	0.77 (0.32, 1.86)	0.602
Left renal	1.03 (0.69, 1.55)	0.888
Center renal	1.13 (0.74, 1.73)	0.548
IMA	1.16 (0.67, 1.99)	0.583
≧1 vessel arising from FL	1.50 (0.98, 2.27)	0.076

^aExcluding those with thrombosed (n = 2) or unknown (n = 60) false lumen (FL) status on initial presenting scan.

CI = confidence interval; COPD = chronic obstructive pulmonary disease; HR = hazard ratio; IMA = inferior mesenteric artery; SMA = superior mesenteric artery.

Table 3.

Outcomes Stratified by Aortic Size (Maximum Descending Aortic Diameter) on Initial Scan

Outcome	<4.5 cm	≧4.5 cm	p
Among all-comers with available imaging (n = 263)
In-hospital mortality	4.8 (6 / 126)	5.8 (8 / 137)	0.698
Aortic intervention	23.8 (30 / 126)	62.0 (85 / 137)	<0.001
Overall mortality	19.0 (24 / 126)	35.8 (49 / 137)	0.003
Failure of medical management	37.3 (47 / 126)	72.3 (99 / 137)	<0.001

Values are percentage (n/N).

Fig 1.

Kaplan-Meier intervention-free survival curve based on aortic size (maximum descending aortic diameter) on initial scan. Data in table presented as % (number at risk) at each time interval.

Comment

Although OMT results in low in-hospital mortality (2.6% to 6.4%) for patients with acute uTBAD, the long-term outcomes remain sobering. Approximately 40% of patients require aortic intervention in the chronic phase, and the intervention-free survival rate at 10 years is 31% [4]. Given these poor long-term outcomes, combined with the well-established ability of endovascular therapy to effectively remodel the aorta in acute complicated TBAD, there has been increased interest in expanding the role of TEVAR to the treatment of acute uTBADs [5, 7, 8]. Ideally, those patients who will ultimately fail OMT could be identified in the early phase of TBAD, when the dissection flap material properties are more conducive to aortic remodeling with TEVAR. In a previous study, we observed that 46% of patients presenting to our institution with acute uTBAD failed OMT and required aortic intervention in the chronic phase [4]. In the present study, we analyzed demographic and radiographic features of this “OMT failure” cohort to identify clinically relevant risk factors for aortic expansion or mortality in the chronic phase of TBAD.

Our analysis of demographic risk factors identified the presence of diabetes mellitus and ESRD to be independent predictors of failure of OMT. These findings may reflect risk factors for patients with reduced life expectancy, especially those patients with renal failure who are on dialysis. It has been recognized that patients with diabetes and/or ESRD may be prone to abnormal aortic calcification and inflammatory changes, however, which could negatively impact patients with aortic dissection [9–11].

The main finding of the current analysis is that a DTA diameter of 4.5 cm or greater at the time of presentation predicts failure of OMT, manifesting in both an increased risk of aortic intervention and reduced long-term survival in patients presenting with acute uTBAD. Contrary to previous reports, the status of the FL did not independently impact aortic intervention or mortality. Furthermore, the number of abdominal visceral vessels originating from the FL had no impact on aortic intervention or survival.

Previous investigations into predictors of aortic growth or mortality in patients with TBAD have identified aortic size and FL status to be important risk factors. In a recent analysis of 254 patients receiving OMT for TBAD, Schwartz and colleagues found that an aortic size of 4.0 cm was predictive of intervention [12]. This finding was in concordance with several other studies that determined that a DTA diameter of greater than 4.0 cm in the acute phase was predictive of both aortic growth and adverse outcomes in the chronic phase of TBAD [13–17]. Our data differ from those previous studies in that a DTA diameter of 4.0 cm in the current study was not predictive of intervention. Instead, we found the size threshold to be 4.5 cm, an aortic diameter that is widely recognized as a significant aneurysm. When patients were divided based on a size of threshold of 4.5 cm and outcomes were analyzed, the results were noteworthy. In patients with aortic diameters greater than or equal to 4.5 cm, aortic intervention was required in 62% of patients, long-term mortality was 36%, and the overall incidence of failure of OMT was 72% (Table 3). These data support the multivariate analysis finding that patients with acute uTBAD who present with an acutely dissected aneurysmal DTA of greater than or equal to 4.5 cm have an increased risk of aortic intervention or death (hazard ratio [HR], 1.39; 95% confidence interval [CI], 1.24 to 1.56; p < 0.001) in the chronic phase when treated exclusively with OMT. They also support the recent findings of Ray and associates identifying a maximum DTA diameter of 4.4 cm to be the most sensitive cutoff for mortality in a stratified analysis of patients presenting with uTBADs [18].

An equally important finding of the current study is that partial FL thrombosis is not an independent predictor of OMT failure in patients with acute uTBAD. These data challenge the findings of Tsai and colleagues who analyzed 201 patients with acute TBAD from the IRAD database with a median follow-up of 2.8 years [19]. In their study, the mean age was 61 years, 69% of patients were men, and the status of the FL was patent in 57% and partially thrombosed in 34% of patients. In that highly influential study, partial FL thrombosis was found to be an independent predictor of mortality (HR, 2.69; 95% CI 1.45 to 4.98; p = 0.002), in patients with acute TBAD. The demographics and radiographic features of the IRAD study population are similar to the cohort studied in the current investigation, with one main difference. Tsai and colleagues included in their study population 55 (27.3%) patients with complicated acute TBAD who underwent either open or endovascular therapy in the acute phase. Complicated TBAD patients, in particular those who undergo open surgery, have a significantly higher in-hospital mortality rate than uTBAD patients, and this undoubtedly impacted the results [20]. Given that TEVAR has become the unequivocal standard of care for acute complicated TBAD, we excluded this cohort of patients from our analysis. Instead we focused only on patients with acute uTBAD, a population in which the optimal therapy has yet to be defined. This appears to be the main difference in the studies and is likely a key contributor to the difference in findings regarding the impact of partial FL thrombosis on mortality. Based upon our univariate analysis, partial FL thrombosis may place patients at higher risk (than patients with patent FL) for the development of late complications, but it is not an independent predictor of OMT failure in patients with acute uTBAD.

The final radiographic feature that was analyzed was the number of visceral vessels arising from the FL. Previous studies have postulated that partial FL thrombosis (depending upon its location) could cause obstruction of distal tears, thus impairing outflow from the FL [19, 21]. This could result in elevated diastolic blood pressures and increased wall tension that, over time, may lead to FL expansion and rupture in the chronic phase. Our own in vitro work in a porcine model of acute TBAD found that distal tear occlusion significantly increased the diastolic component of FL pressure [22]. Therefore, all aortic branches (visceral, lumbar, and intercostal arteries) originating from the FL may be considered to represent outflow from the FL. By this rationale, patients with zero or 1 visceral vessel originating from the FL should be at increased risk for aortic expansion and/or failure of OMT, compared with patients with multiple vessels originating from the FL. Nevertheless, the findings of the univariate (HR, 1.50; 95% CI, 0.98 to 2.27) analysis demonstrated no significant impact of the number of visceral vessels originating from the FL upon the need for aortic intervention or mortality (Table 2).

Major limitations of the current study include the lack of complete imaging data, an incomplete knowledge of the details and efficacy of the antihypertensive regimens, and the lack of complete aortic-related mortality data on follow-up. Although we were able to analyze aortic size, FL status, and true or FL status of visceral vessels in 81% of patients, the missing data from 19% of patients could have influenced our results. The absence of blood pressure data limited our ability to assess the efficacy of the antihypertensive regimens prescribed in patients receiving OMT. Inadequate blood pressure control and antiimpulse therapy could have significantly impacted the efficacy of OMT in preventing aortic expansion and aortic-related mortality. Finally, we acknowledge that our understanding of the success of OMT in the treatment of patients with uTBAD would be greatly enhanced with aortic-related mortality data. Unfortunately, these data were unobtainable, as many patients changed addresses or phone numbers, or died outside our hospital system. Thus, we could not perform subgroup analyses among our study cohort to assess the role of any potentially confounding comorbidities on aortic-related pathology and outcomes.

The current study represents one of the largest populations of uTBAD patients to be analyzed and serves as a continuation of our previous work regarding the impact of endovascular therapy on long-term outcomes of TBAD [4]. As mentioned previously, in order to advance the treatment algorithm of TBAD, an improved understanding of the pathophysiology is necessary. Over a decade of experience of using TEVAR in the treatment of patients with complicated acute TBAD has led to a paradigm change with this high-risk variant of TBAD. The current study adds to the growing body of data that demonstrates the inadequacy of medical therapy for uTBAD. The next step will to be to develop a highly accurate predictive model to determine which patients will benefit from early endovascular therapy to remodel the aorta and prevent late complications. We also advocate for the formation of a multidisciplinary group to develop guidelines regarding the contemporary management of uTBAD. Our data suggest that at the time of diagnosis, uTBAD patients with diabetes mellitus, ESRD, dissections that involve the thoracic and abdominal aorta, and a descending thoracic aortic diameter of 4.5 cm or greater are at high risk for aortic expansion and late mortality. These patients should undergo a more aggressive surveillance imaging protocol and may benefit from TEVAR, in addition to OMT, at the index hospitalization.