Contemporary results of open repair of ruptured descending thoracic and thoracoabdominal aortic aneurysms

Results 

The baseline characteristics of the patients in our study are listed in Table I. The extent of the aneurysms was as follows: type I in 8 patients, type II in 3, type III in 11, type IV in 13, and descending thoracic in 6. The time from the onset of symptoms to operation was less than 6 hours in 10 patients, between 6 and 24 hours in 13 patients, and more than 24 hours in 16 patients; this information was not available in 2 patients. The mode of transportation at presentation was by air in 28 cases (68.3%) and by ground transport in 9. Four patients were inpatients at the time of their ruptures; all patients were admitted with symptoms consistent with but had computed tomographic scans that did not demonstrate rupture. They subsequently all had clinical deterioration (three with hypotension) and intraoperative findings consistent with rupture. Those patients with a shorter duration from symptom onset to operative repair were more likely to present in shock requiring CPR with an in-hospital mortality of 50% vs a mortality of 6.7% (P < .05) in those with symptom duration of greater than 24 hours (Appendix Table I, Appendix Table II). Otherwise, the preoperative characteristics and postoperative outcomes were similar between these groups.

Table I. Baseline characteristics of patients undergoing repair of ruptured DTA and TAAA

	Variable	Data
	Age (y)	75.5±6.6
	Female	39.0(16)
	White	94.9(39)
	History of COPD	45.0(18)
	History of CVA	15.0(6)
	History of CAD	67.5(27)
	Smoking	38.5(15)
	History of diabetes mellitus	2.5(1)
	Any prior aortic reconstruction	36.6(15)
	Preoperative characteristics
	Symptom onset to OR (h)
	<6	25.6(10)
	6-24	35.9(14)
	>24	38.5(15)
	Mean time	44.9±57.5
	Preoperative shock	37.5(15)
	Preoperative CPR	17.1(7)
	Aneurysm size and blood chemistry
	Mean aneurysm diameter (cm)	8.4±2.3
	Free rupture	26.8(11)
	Mycotic	7.3(3)
	Mean BUN (mg/dL)	24.3±10.8
	Mean creatinine (mg/dL)	1.4±0.7
	Mean hemoglobin (g/dL)	11.7±4.3
	Operative characteristics
	After 2002	53.7(22)
	OR time (h)	4.4±1.7
	OR fluid used	4598±2262
	Blood units used	7.8±5.7
	Lowest SBP (mm Hg)	72.2±21.7
	Intraoperative CPR	21.9(9)
	Sequential clamp use	24.3(9)
	Clamp time (min)	65.1±19.5
	Visceral ischemia time (min)	44.3±15.2
	Renal ischemia time (min)	43.1±23.9
	Renal cold perfusion	37.1(13)
	Renal bypass	28.6(10)
	Intercostal reimplantation	2.9(1)
	Left heart bypass	41.0(16)
	CSF drain	55.0(22)
	Naloxone	22.5(9)

Data are % (n) or mean ± SD. COPD, Chronic obstructive pulmonary disease; CVA, cerebrovascular accident; CAD, coronary artery disease; OR, operating room; CPR, cardiopulmonary resuscitation; BUN, blood urea nitrogen; SBP, systolic blood pressure; CSF, cerebrospinal fluid.

The mean duration of follow-up for those who survived to hospital discharge was 28.4 ± 26.6 months (range, 1.4-85.6 months). There was one postoperative stroke; this patient died at day 7 after repair. Of the patients surviving to discharge, only 6 patients were discharged home, whereas 24 were discharged either to a skilled nursing facility or rehabilitation center. Four patients required permanent dialysis, one of whom became a long-term survivor.

The operative mortality rate was 26.8% (11/41). All but two deaths occurred within 24 hours of operation. The intraoperative mortality rate was 17% (7/41); 5 of these patients had received preoperative CPR. Four deaths occurred during dissection before the graft could even be implanted. Two other deaths were due to hemorrhagic shock and cardiac arrest shortly after operation. After 24 hours, one death occurred at postoperative day 7 and another at day 31, when the family withdrew support. After hospital discharge, eight additional patients died within 6 months of discharge; the mortality rate at 1 year was 46.3%. Two-year survival curves are represented in Fig 1. The remaining deaths were late deaths, occurring after at least 1 year of survival. Although the cause of death could not be determined in two patients, none of the other deaths was aneurysm related. Factors associated with in-hospital mortality included intraoperative hypotension (P = .007) and a shorter duration from symptom onset to operation (Table II; P = .04). The presence of chronic obstructive pulmonary disease, an increased preoperative blood urea nitrogen, and the presence of type II and III aneurysms were associated with increased mortality after discharge from the hospital (Table III; P < .05).

View Large Image Download to PowerPoint

Fig 1. Cumulative survival of patients undergoing surgical repair of ruptured descending thoracic and thoracoabdominal aneurysms.

Table II. Univariate analysis of variables tested for association with in-hospital mortality

	Variable	In-hospital mortality		P value
		No (n = 30)	Yes (n = 11)
	Age (y)	75.5±6.6	75.5±6.8	.93
	History of COPD	45.0(12)	60.0(6)	.30
	History of CVA	13.3(4)	20.0(2)	.63
	History of CAD	63.3(19)	80.0(8)	.45
	Smoking	40.0(12)	33.3(3)	1.0
	Any prior aortic reconstruction	36.7(11)	36.7(4)	1.0
	Preoperative characteristics
	Symptom onset to OR (h)			.041
	<6	17.2(5)	50.0(5)
	6-24	34.5(10)	40.0(4)
	>24	48.3(14)	10.0(1)
	Mean time	53.9±62.5	18.7±28.3	.07
	Preoperative shock	26.7(8)	70.0(7)	.024
	Preoperative CPR	6.7(2)	45.4(5)	.01
	Aneurysm size and blood chemistry
	Aneurysm diameter (cm)	8.1±1.9	9.2±3.0	.33
	Free rupture	16.7(5)	54.5(6)	.041
	Mycotic	3.3(1)	18.2(2)	.17
	BUN	27.7±11.1	23.1±10.3	.66
	Creatinine	1.4±0.8	1.3±0.5	.85
	Hemoglobin	11.3±2.2	12.8±7.6	.92
	Operative characteristics
	OR time (h)	4.9±1.2	3.3±2.4	.08
	Excluding intraoperative deaths		5.2±1.4	.69
	OR fluid used	4814±2275	3902±2199	.21
	Blood units used	6.4±3.8	12.7±8.2	.033
	Lowest SBP	78.8±13.4	50.3±29.6	.007
	Operative CPR	3.3(1)	72.3(8)	<.001
	Sequential clamping	24.1(7)	25.0(2)	1.0
	Clamp time	63.4±18.9	76.7±22.3	.28
	Visceral ischemia time	42.5±15.1	54.2±12.7	.14
	Renal ischemia time	38.8±22.4	71.2±11.1	.014
	Renal cold perfusion	30.0(9)	80.0(4)	.052
	Renal bypass	26.7(8)	40.0(2)	.61
	Intercostal reimplantation	3.3(1)	0(0)	1.0
	Left heart bypass	46.7(14)	22.2(2)	.26
	CSF drain	70.0(21)	10.0(1)	.002
	Narcan	26.7(8)	10.0(1)	.40
	Outcomes
	Dialysis (any)	20.7(6)	0(0)	1.0
	Tracheostomy	33.3(10)	100.0(1)	.35
	Pneumonia	40.0(12)	100.0(2)	.18
	Paraplegia/paraparesis	13.3(4)	50.0(1)	.29

Data are % (n) or mean±SD. COPD, Chronic obstructive pulmonary disease; CVA, cerebrovascular accident; CAD, coronary artery disease; OR, operating room; CPR, cardiopulmonary resuscitation; BUN, blood urea nitrogen; SBP, systolic blood pressure; CSF, cerebrospinal fluid.

Table III. Factors associated with mortality within the first 2 years after discharge from the hospital (Cox proportional hazards methodology)

	Variable	Hazard ratio	95% Confidence interval	P value
	COPD	17.68	1.47-213.28	.024
	BUN	1.12	1.03-1.21	.005
	Extent of repair (vs type IV)
	Type I	1.32	0.10-17.49	.83
	Type II	39.28	1.41-1095.27	.031
	Type III	101.73	3.62-2856.82	.007
	Isolated DTA	3.51	0.55-22.58	.18

COPD, Chronic obstructive pulmonary disease; BUN, blood urea nitrogen; DTA, descending thoracic aneurysm.

There were 12 octogenarians. Three patients did not survive to hospital discharge, accounting for an operative mortality of 25% as compared with 27.6% in those younger than 80 years (not significant). Three additional patients died within 4 months of hospital discharge, resulting in a 1-year mortality rate of 50%. One additional patient died 22 months later, 1 month before his 90th birthday. The remaining five survivors are alive and well at a mean of 24 months (range, 10-34 months) follow-up. The baseline characteristics and comorbidities, except for age, did not differ between the groups. The only differences observed were lower mean hemoglobin (9.6 ± 1.9 g/dL vs 12.6 ± 4.8 g/dL; P = .005) and shorter mean renal ischemia time (22.4 ± 23.2 minutes vs 49.4 ± 20.7 minutes; P = .02) in those 80 years or older as compared with the younger cohort. Kaplan-Meier survival estimates (Fig 2, A) at 1 and 2 years were 50% and 37.5%, respectively, for octogenarians, whereas the corresponding figures were 55.2% and 50.6% for the younger patients (P = .72).

View Large Image Download to PowerPoint

Fig 2. Kaplan-Meier analysis of survival after descending thoracic aneurysm and thoracoabdominal aortic aneurysm repair. Survival is analyzed according to age at operation (A) and year of operation (B).

Spinal cord ischemia developed in five (15.6%) of those who survived the operation for more than 24 hours and allowed neurologic evaluation: paraplegia in four patients and paraparesis in one additional patient. The lone patient who developed paraparesis recovered from it without deficits. Paraplegia was delayed in onset in three of four patients, all cases of which were associated with postoperative hypotension. Of the four patients who developed paraplegia, two had intraoperative CSF drainage, and both had delayed paraplegia develop: one while the drain was still in place and the other after its removal. The latter had the drainage reinstituted without benefit. The remaining two patients with paraplegia did not have intraoperative CSF drainage. One of these patients experienced delayed paraplegia and had a lumbar drain placed; no improvement in neurologic function was achieved. Risk factors associated with spinal cord ischemia were extent of the aneurysm, with type II aneurysms having the highest risk for the development of paraplegia (P < .0001; Table IV), and increased aortic cross-clamp time (P = .03).

Table IV. Factors associated with increased morbidity and mortality with logistic regression analysis

	Outcome	Odds ratio	95% Confidence interval	P value
	Death
	Lowest OR blood pressure	0.9	0.82-0.99	.037
	Need for tracheostomy
	Increased aneurysm size (each 1 cm)	5.4	1.12-26.48	.036
	COPD	26.0	1.57-431.1	.023
	Paraplegia
	Extent of aneurysm⁎			<.001
	Type II vs I	2.0
	Type II vs III	1.3
	Type II vs IV	2.0
	Type II vs isolated DTA	2.0
	Clamp time	1.1	1.01-1.16	.03
	Dialysis
	Increased baseline creatinine	31.2	1.6-591.1	.02

OR, Operating room; COPD, chronic obstructive pulmonary disease; DTA, descending thoracic aortic aneurysm.

⁎ Analysis of variance.

Pulmonary failure requiring tracheostomy developed in 35.5% of patients. Controlling for other operative and patient characteristics, chronic obstructive pulmonary disease (P = .023) and a larger aneurysm were predictive of the need for tracheostomy. For each increase in size of 1 cm, there was more than a fivefold risk for prolonged ventilatory support and progression to tracheostomy (odds ratio, 5.44; P = .036). The incidence of renal failure necessitating either temporary (n = 3) or permanent (n = 4) dialysis was 19.3%. The only identified predictor was increased baseline creatinine (P = .02).

The mortality rate during the first half (19 of 41 cases from 1995 through 2002) of this series was 42.1%, as compared with 13.6% in the latter half of our experience (22 of 41 cases since 2003). Although this did not reach statistical significance, there was a strong trend toward improved survival (Fig 2, B; P = .075). There were no differences in the baseline patient characteristics or preoperative variables (including preoperative shock and the need for CPR) over time, whereas there were significant differences in the intraoperative variables; patients in the latter half of this series had significantly fewer blood transfusions (10.4 vs 5.9 units; P = .037) and a shorter period of renal ischemia (53.5 vs 35.2 minutes; P = .043).

Discussion 

This study illustrates that repair of RDTAA still poses a daunting challenge to both the surgeon and the patient. It is a highly lethal problem with lasting effects. According to one report, approximately 50% of patients with rupture of a thoracic aortic aneurysm die within 6 hours of the onset of symptoms, and an additional 25% die between 6 and 24 hours, thus allowing little time for surgical repair.11 In this series, the mortality rate was 26.8%, with most deaths (82%) occurring within the first 24 hours after operation and well over half (62%) of late deaths taking place within 4 months of discharge from hospital, thus reflecting a sustained increased risk of mortality even after successful repair and hospital discharge. This corroborates the findings of earlier studies that there is increased age-adjusted mortality throughout the first postoperative year after repair of both ruptured TAAAs and abdominal aortic aneurysms.3, 12 The overall actuarial survival at 1 year was 63.6% and at 2 years was 48.5%. For the hospital survivors, the corresponding figures were 73.3% and 64.4%.

The reason for the high morbidity and mortality is multifactorial and clearly more complex than the hemodynamic instability of the patient at presentation. Previous investigations have shown that advanced age, female sex, type II aneurysms, preoperative renal insufficiency, increased aortic clamp time, preoperative shock, coronary events, and low hospital and surgeon volume increase the risk of adverse outcomes after repair of ruptured or emergent repair of DTA and TAAA.1, 3, 13, 14, 15, 16, 17 Our data confirm that preoperative shock and type II aneurysm portend an increased incidence of mortality and paraplegia, respectively.

A number of earlier reports are difficult to interpret because of the heterogeneous nature of their clinical material. Some included aortic dissections,13, 18, 19, 20, 21 whereas others incorporated symptomatic aneurysms without rupture in their reviews.15, 16 Furthermore, reporting standards for mortality are not uniform across the literature; some report 30-day mortality, and others report in-hospital mortality. The results in this series are within the reported range in the available literature. In a review of statewide experience, Rigberg et al3 reported a 30-day mortality rate of 48.4%, a finding similar to the 53.8% rate in a survey of the Nationwide Inpatient Sample reported by Cowan and associates.1 Smaller institutional series2, 13, 15, 16, 17, 19 have reported mortality rates ranging from 15% to 67%. The mid-term survival rates seen in this series are also in accordance with the findings of earlier reports3, 16, 22 in which 1-year mortality rates have been reported between 47% and 61.5%.

With respect to the effects of age on mortality, a progressive decrease in 1-year survival rates with advanced age, from 44.4% for patients in their 50s to 31.2% for those in their 80s, has been noted.3 Huynh and colleagues23 reported a 50% 30-day mortality rate for “high-risk” octogenarians presenting with rupture, diabetes, or congestive heart failure. In this regard, the results in this report compare favorably; the 1-year survival estimate for octogenarians was 50%, and the 2-year estimate was 37.5%. In our experience, nearly 30% of cases (12/41) were octogenarians, and their mortality did not differ from that of the younger patients. This occurred even though the octogenarians were more anemic at presentation than the younger population; it is of note, however, that they also had a shorter renal ischemic time. Although advanced age has been found to be a predictor of poor acute and long-term outcome,1, 3, 23 our data indicate that age alone is not a major criterion in the decision-making process of whether to offer repair of RDTAAs.

Although adjunctive end-organ protection methods have effectively reduced adverse outcomes in elective settings,24, 25, 26 the time constraints encountered in the life-threatening condition of RDTAA dissuade their routine use. Girardi et al17 used CSF drainage in only 50% of their patients and reported a paraplegia rate of only 5%, similar to their elective data. LeMaire et al19 failed to demonstrate beneficial effects of left heart bypass or intercostal artery attachment against the development of paraplegia in the absence of CSF drainage. In the present study, the efficacy of distal aortic perfusion, CSF drainage, or naloxone could not be demonstrated, largely because it was not routinely used and because relatively few patients had events. The incidence rate of spinal cord ischemia is similar to the rates published in other series of ruptured TAAAs, which almost universally are significantly higher than those seen in elective cases.16, 27 Given the low numbers and lack of randomization, the efficacy of adjunctive measures will likely need to be extrapolated from elective series.

Associations with outcome and hospital volume are somewhat controversial, because some authors found that hospital experience did result in an improvement in mortality in the setting of elective thoracic aneurysm repair.3, 28, 29 In the emergent setting, such an association has not been uniformly observed. Schepens et al22 reported no improvement in mortality with time; Rigberg et al3 found that only low-volume centers (one case per year) were associated with increased mortality at 1 year. The observed superior outcome in the centers of excellence may be in part related to intrinsic factors related to patient presentation and referral patterns. Crawford and colleagues’18 original report of 117 ruptured thoracic and thoracoabdominal aneurysms described an onset of symptoms related to rupture of less than 24 hours in only 16% of their patients, and as many as 34% of patients had symptoms exceeding 1 week. This may represent a selection or referral bias favoring outcomes, because our data would suggest that patients undergoing operation sooner after symptom onset have the highest mortality. In a report by LeMaire and associates,19 97% of their patients were hemodynamically stable at presentation. This is in contrast to the present series, in which 36.6% of patients presented with a systolic blood pressure less than 90 mm Hg. We made every effort to accept all patients from outlying facilities with an intention to repair, unless they were deemed medically inoperable. This is well reflected by the high number of intraoperative (7/11) deaths and deaths within the first 24 hours. These patients were almost universally in extremis at presentation.

Compromised functional status after urgent/emergent TAAA repair has been detailed in a report by Rectenwald et al.16 In that study, a “bad” outcome—defined as nonambulatory status, discharge to a long-term care facility, or death—was noted in 44% of patients surviving to discharge. However, at 1 year, 87% of patients with follow-up were at home and ambulatory. Similarly, only 6 of 30 surviving patients in this series were discharged home, whereas 24 were discharged either to a skilled nursing facility or to a rehabilitation center. Only two of these were in a long-term nursing home at last follow-up, including the lone long-term (3 years) survivor with paraplegia. Despite the high demand for extended care even after hospital discharge, repair of RDTAA seems well justified because it is a uniformly lethal condition without therapy.

The data in this series strongly suggest the pertinence of volume regarding outcome. Our volume has increased in recent years such that the half of our experience was accumulated over the past 3.5 years and has been concentrated on two surgeons. Although it did not reach significance, owing to small number of patients, a strong trend toward improved mortality in the latter part of our experience most likely represents improved efficiency in the management of these patients with enhanced surgeon and institutional experience. This trend was associated with a more conservative approach to fluid management, with approximately half the intraoperative blood transfusion in the second half of our experience despite similar degrees of hemodynamic instability and patient comorbidities. Thus, prompt diagnosis and expeditious intervention or referral to high-volume centers would be desirable to maximize the chance of optimal outcome.

High incidences of adverse outcomes and laborious, complicated postoperative courses have driven the need for alternative methods of repair. With successful application of endovascular technology in the arena of isolated DTA,30 its horizon is being expanded to include treatment of thoracoabdominal and arch lesions using branched graft technology4, 5, 6 and hybrid procedures,7, 8, 9 with promising preliminary results. In a series of 26 TAAAs treated by hybrid procedures, no paraplegia and a mortality rate of 13% were noted in the 23 elective patients, whereas all 3 patients who presented with rupture died.7 In another series, an overall mortality rate of 30% and paraplegia rate of 20% were reported; in two patients who presented with rupture, there was one paraplegia with no death.9 The results with these alternative approaches may improve with further refinement of technology and skills and may be applicable to repair of ruptured aneurysms. Until then, direct surgical repair for RDTAA provides a meaningful way of saving lives. There is still room for improvement, however, and concerted efforts need to be made to improve surgical outcome. Regionalization of this complex aortic surgery to centers of excellence with high volume should be considered. This series contributes to contemporary results of open repair of ruptured thoracic and thoracoabdominal aneurysms and helps provide a standard against which results of future endovascular endeavors should be held.

Author contributions