Recent risk factors for open surgical mortality in patients with ruptured abdominal aortic aneurysm

Abstract

Aim

We examined recent relevant prognostic factors for the outcome of open surgical treatment of ruptured abdominal aortic aneurysm.

Methods

Between 2006 and 2012, 35 patients received emergency open surgical treatment for ruptured abdominal aortic aneurysm at our institute. We reviewed ambulance activity logs and clinical records of 34 infrarenal ruptured abdominal aortic aneurysm patients retrospectively. Univariate and multivariate logistic regression analyses were carried out to identify risk factors for surgical outcomes.

Results

Eight patients died during surgery or within a few hours following surgery completion. Through univariate analysis, body mass index, serum lactate level, arterial blood pH, base excess, platelet count, prothrombin time–international normalized ratio, activated partial thromboplastin time, type of ruptured aneurysm, response to i.v. fluid resuscitation within 2,000 mL in the initial therapy, and volume of blood loss during surgery were detected to be significant variants. Multivariate logistic regression analysis revealed the patients who were hemodynamically stabilized after primary volume loading had a 13.2 times higher possibility of survival. Body mass index, high serum lactate level, and volume of blood loss were also found to be independent risk factors of mortality.

Conclusion

The risk factors of open surgical ruptured abdominal aortic aneurysm repair, body mass index, lactate level, volume of intraoperative blood loss, and response to initial 2,000 mL fluid resuscitation were correlated to survival.

Introduction

Recently, the elective operative repair of abdominal aortic aneurysm (AAA) has been associated with low morbidity and mortality in properly selected individuals.1, 2, 3 However, the mortality of ruptured AAA (rAAA) remains high despite advances in diagnostic methods, techniques for repair including endovascular aneurysm repair (EVAR), and intensive care.1

Previously, several factors such as age, female sex, prolonged collapsed vital phase, massive transfusion, or hemorrhage during surgery, were considered independent risk factors of mortality.2, 3, 4, 5, 6, 7, 8, 9, 10, 11 However, these results are still not definitive and some studies were published in the 1990s.

This study aimed to define relevant prognostic factors for outcomes of open surgical rAAA treatment.

Methods

Our cardiovascular team consisted of nine surgeons who perform approximately 30–40 elective AAA, including EVAR, and 5–10 emergent AAA repairs per year. No in‐hospital or early postoperative death is observed for elective AAA repair.

Between 2006 and 2012, 57 patients received emergency surgical treatment for AAA. Twenty‐two patients were excluded from the study because aneurysms were impending rupture type; one other patient was excluded because surgical management for suprarenal AAA was largely different. The remaining 34 patients were the subject of this study. Before surgical management, we attempted to maintain systolic blood pressure in the range of 60–90 mmHg to avoid reduction of cerebral blood flow and secondary hypertension induced by excessive fluid resuscitation.

The patients were divided into survivors who were conclusively discharged to home or transferred for rehabilitation, or non‐survivors who died during or after surgery. The factors previously reported or additional factors assumed to possibly influence the outcomes were compared.

Preoperative status

Demographics such as sex, age, body mass index (BMI), medical history, and previous abdominal surgery, were investigated from the clinical records.

In addition to blood analysis at hospital arrival, we investigated the patients' hemodynamic status prior to surgery, including the need for cardiopulmonary resuscitation (CPR), response to i.v. fluid resuscitation, duration from symptom onset to hospital arrival, and from hospital arrival to skin incision. These data were collected from ambulance activity logs and clinical records. The response to i.v. fluid resuscitation was evaluated by rapid i.v. infusion of within 2,000 mL crystalloid or transfusion of blood type O. Patients whose hemodynamics improved and stabilized prior to surgery by initial 2,000 mL volume loading were defined as responders, and those without improvement were classified as non‐responders.

Operative data

We evaluated variants such as volume of blood loss during surgery including hematoma, operative time, and volume of intraoperative blood transfusion. We also investigated clamping of the proximal aorta before cardiopulmonary collapse, duration from skin incision to application of the aortic clamp, and aortic clamp site.

Type of ruptured aneurysm

Computed tomography imaging was carried out in all patients prior to surgery. The ruptured aneurysm was classified according to the Fitzgerald classification.12

Statistical analysis

Univariate analysis was carried out between the two groups. The data with a normal distribution were compared using an unpaired Student's t‐test. Non‐parametric numerical data were compared using the Mann–Whitney U‐test. Fisher's exact test was used for the categorical data.

Multivariate forward stepwise logistic regression analysis was used to determine independent predictors of mortality. All variables with a P‐value less than 0.05 in the univariate analysis were selected, and forward stepwise regression was carried out. A value of P < 0.05 was considered statistically significant. All statistical analyses were carried out using spss 15.0 (SPSS, Chicago, IL, USA).

Results

There were no hospital deaths in the survivors group and almost all patients were followed up by our hospital. Eight patients died during surgery or a few hours after (Table 1). Six patients died of hemodynamic deterioration during surgery. One patient with collapsed hemodynamics required CPR on the way to our hospital; in this patient, extensive necrosis of the colon was confirmed during surgery. The other patient showed ventricular fibrillation shortly after proximal anastomosis and declamping of the aorta; in this patient, the serum potassium level was 4.2 mEq/L and ST depression was observed with a change in rhythm. This was supposed not to be the result of reperfusion syndrome but the result of acute coronary syndrome, which might be induced by low coronary perfusion. Of the five patients who died prior to aortic proximal clamping, four were diagnosed as rAAA in other small hospitals and transferred long distances for surgery with only peripheral blood access. These patients were moved quickly to the operating room, but their surgical conditions were already difficult. One patient died because of uncontrollable bleeding induced by disseminated intravascular coagulation.

Table 1.

Characteristics of non‐survivors of emergency open surgical treatment for ruptured abdominal aortic aneurysm (n = 8)

Case	Age, years	Sex	Fitzgerald classification of ruptured aneurysm	Ao clamp	Cause of death
1	64	Male	3	×	Bleeding
2	79	Male	4	×	Bleeding/intestinal necrosis
3	76	Male	3	○	VF after declamping
4	73	Male	3	×	Bleeding
5	74	Male	3	×	Bleeding
6	62	Male	4	○	Bleeding
7	79	Male	4	×	Bleeding
8	93	Male	4	○	Bleeding/DIC

×, unable to aortic proximal clamping during surgery; ○, able to aortic proximal clamping during surgery; Ao, proximal aorta; DIC, disseminated intravascular coagulation; VF, ventricular fibrillation.

Preoperative status

No significant differences except BMI were noted from the standpoint of demographics (Table 2). Preoperative biochemical data did not differ between the two groups (Table 3). The platelet count was significantly higher and the lactate level was significantly lower in survivors than in non‐survivors. In addition, there were significant differences in arterial blood pH level, base excess, prothrombin time–international normalized ratio, and activated partial thromboplastin time.

Table 2.

Demographic data of patients who underwent emergency open surgical treatment for ruptured abdominal aortic aneurysm (n = 34)

	Survivors	Non‐survivors	P‐value
Total number	26	8
Sex, male (%)	21 (80.1%)	8 (100%)	0.31
Age, years	74.7 ± 8.4	75.0 ± 9.0	0.94
BMI	21.8 ± 3.1	25.9 ± 2.5	0.002
Smoker	20 (76.9%)	7 (87.6%)	0.66
Medical history
Hypertension	26 (100%)	8 (100%)	1.00
Ischemic heart disease	3 (12%)	3 (37.5%)	0.13
Pulmonary disease	3 (12%)	2 (25%)	0.57
Previous abdominal operation	4 (15.4%)	2 (25%)	0.61
Onset to arrival time (min)	123.5 ± 101.9	236.1 ± 388.9	0.44
Arrival to skin incision time (min)	89.9 ± 48.1	66.0 ± 30.8	0.21

BMI, body mass index.

Table 3.

Hematological examination in patients who underwent emergency open surgical treatment for ruptured abdominal aortic aneurysm (n = 34)

	Survivors	Non‐survivors	P‐value
HgB (g/dL)	10.8 ± 2.1	9.2 ± 2.1	0.071
Plt (104/μL)	17.7 ± 4.9	14.1 ± 7.8	0.029
AST (IU/L)	24.3 ± 10.5	24.8 ± 8.1	0.91
Cr (g/dL)	1.37 ± 0.51	1.49 ± 0.81	0.66
CK (IU/L)	101.3 ± 55.9	135.5 ± 93.8	0.22
pH	7.34 ± 0.06	7.09 ± 0.25	0.006
B.E.	−5.09 ± 4.3	−13.1 ± 8.7	0.002
Lactate (mg/dL)	46.3 ± 27.4	92.4 ± 36.8	0.008
PT‐INR	1.13 ± 0.22	1.47 ± 0.41	0.028
APTT (seconds)	33.5 ± 35.1	52.4 ± 56.1	0.028
Fibrinogen (mg/dL)	311.2 ± 103.2	233.5 ± 120.3	0.09

APTT, activated partial thromboplastin time; AST, aspartate aminotransferase; B.E., base excess; CK, creatine kinase; Cr, creatinine; HgB, hemoglobin; Plt, platelet count; PT‐INR, prothrombin time–international normalized ratio.

Although all patients experienced shock (systolic blood pressure <90 mmHg) at least once before surgery, the patients in the survivors group stabilized hemodynamically after volume loading (Table 4). One patient in the survivors group required 2 min of CPR after sudden loss of consciousness and circulatory collapse as a result of massive bleeding in the emergency room.

Table 4.

Preoperative factors in patients who underwent emergency open surgical treatment for ruptured abdominal aortic aneurysm (n = 34)

	Survivors	Non‐survivors	P‐value
Total number	26	8
Response to volume load			0.0001
Responders	25 (96.2%)	2 (25.0%)
Non‐responders	1 (3.8%)	6 (75.0%)
CPA before skin incision	1 (3.7%)	4 (50.0%)	0.007
IABO use	6 (23.1%)	3 (37.5%)	0.65
Fitzgerald category, mortality (%)			0.03
F‐1 or F‐2	12	0 (0.0%)
F‐3 or F‐4	14	8 (34.8%)

CPA, cardiopulmonary arrest; IABO, intra‐aortic balloon occlusion; Non‐responders, patients not hemodynamically improved by volume loading; Responders, patients hemodynamically improved and stabilized by volume loading.

The Fitzgerald classification of all cases is shown in Table 4. All F‐1 or F‐2 classified patients survived. The mortality of patients in the F‐3 or F‐4 category was 34.8%. The aneurysm classification was significantly different between the two groups (P = 0.03).

Operative data

The volume of intraoperative blood loss, blood transfusion, surgical time, duration time from skin incision to aortic clamping, and selection of the clamp site were comparable between the groups (Table 5). We found the volume of blood loss also showed significant difference.

Table 5.

Operative data of patients who underwent emergency open surgical treatment for ruptured abdominal aortic aneurysm (n = 34)

	Survivors	Non‐survivors	P‐value
Total number	26	8
Skin incision to Ao clamp time (min)	53.0 ± 21.4	95.9 ± 83.9	0.41
Unable to Ao clamp	0 (0%)	5 (62.5%)	0.0002
Clamp site			0.99
Suprarenal	3 (11.5%)	0 (0%)
Infrarenal	23 (88.5%)	3 (100%)
BT volume (mL)	6,096.2 ± 3,188	8,855.0 ± 5,969	0.12
Surgical time (min)	333.1 ± 127.6	186.9 ± 119.5
Blood loss volume (mL)	5,058.9 ± 3,790	9,803.6 ± 7,785	0.03

Ao, proximal aorta; BT, blood transfusion.

We could not access a suitable surgical clamp site before hemodynamic collapse in five patients mainly because of preoperative uncontrollable vitals, massive hematoma, fat, or intestinal edema.

Multivariate analysis

The univariate analysis revealed BMI, responder status, blood lactate level, base excess, arterial blood pH level, platelet count, activated partial thromboplastin time, prothrombin time–international normalized ratio, the Fitzgerald classification, and volume of intraoperative blood loss as significant variants.

The multivariate analysis revealed that responders following primary volume load within 2,000 mL had significant differences. The possibility of survival of responder patients was estimated to be 13.2 times higher than that of non‐responders.

Body mass index, high serum lactate level, and volume of blood loss were also found to be independent risk factors of surgical death (Table 6).

Table 6.

Predictors for survival in patients with ruptured abdominal aortic aneurysm

Multivariate analysis	Odds ratio	95% CI	P‐value
BMI	0.39	0.26–0.67	0.03
Responders	13.20	5.73–1,720.10	0.007
Lactate	0.27	0.011–0.590	0.02
B.E.	0.76	0.61–1.35	0.09
APTT	0.63	0.35–1.04	0.07
Blood loss volume (mL)	0.41	0.27–0.65	0.04

APTT, activated partial thromboplastin time; B.E., base excess; BMI, body mass index; Responders, patients hemodynamically improved and stabilized by volume loading.

Discussion

The mortality of emergency surgery including EVAR for rAAA is still high and is reported to be 40–70% even if the patients were alive on arrival at the hospital.2, 3, 4

Although many attempts to evaluate the risk factors of mortality in patients with rAAA have been published, no systems or variables have proven to be reliable in the accurate prediction of mortality.5

Previous studies have reported that preoperative systolic blood pressure <90 mmHg, prolonged shock (e.g., >5 h), unconsciousness, and cardiac arrest are independent risk factors of mortality.13, 14, 15 Our results also showed that hemodynamic failure was one of the most important risk factors of rAAA. Furthermore, almost all survivors responded and stabilized after initial volume loading. Dick and colleagues reported excessive volume resuscitation induced hemodilution, resulting in coagulation disorder and increased perioperative risk of death.16 Volume loading should be controlled to the minimum amount, but also at a level necessary for systolic blood pressure to maintain cerebral perfusion. If the patients' hemodynamics have not stabilized, there is no reason to limit volume resuscitation and quick decisions need to be made for commencing emergency laparotomy or thoracotomy in the emergency unit. In our institute, we choose to move to the operation unit promptly because staff and materials are already well prepared and less time is wasted.

Although age of 80–85 years was considered a risk factor of mortality,2, 15 we found no significant difference with respect to age, as Biancari et al. reported.6, 7, 17 Although the elderly patients have an increased risk of having cardiac and cerebrovascular events in the postoperative period,8 the risk factor might not be age itself but cardiac or some other capacities of these patients.9 Regarding the blood data, previous studies reported hemoglobin or hematocrit levels indicated the degree of hemorrhage, and elevated serum creatinine level >1.5 mg/dL correlated with adverse outcomes.10 However, we found no correlation between these values and outcomes.

Even as elevated initial and 24‐h serum lactate levels are reported to be significantly correlated with mortality of surgical intensive care unit patients,11, 18 immediate postoperative serum lactate level <36 mg/dL is shown to be good predictors of rAAA.19 Our study also showed the blood lactate level was an independent risk factor of mortality.

The most suitable and accessible clamp site of the aorta needs to be discerned immediately, including use of an occlusion balloon or descending aortic clamp with the left lateral thoracotomy. One of the well‐known classifications of rAAA from the viewpoint of ongoing hemorrhage is the Fitzgerald classification. This classification indicates not only the progression of a hemorrhage but also strategic selection of how to treat and manage massive bleeding. In this study, non‐survivors all belonged to category F‐3 or F‐4.

The relation of high BMI and high mortality rate in patients with rAAA has never been reported and is unclear. We considered that these patients needed a longer time for laparotomy to be carried out, to clear the surgical field, and to clamp the aorta, thus resulting in one of the independent risk factors. Wakefield et al. described a surgical time of >400 min correlated with higher mortality, and other previous studies showed excessive hemorrhage of >11,000 mL and transfusion of >6,000–7,000 mL were independent risk factors of intraoperative death.2, 9, 14, 20, 21 We found blood transfusion for non‐survivors was more than 8,000 mL on average, but this was not a significant difference compared with that of survivors. As almost all non‐survivors' operations were not accomplished, we did not compare the surgical time statistically. The volume of intraoperative blood loss showed significant difference between the two groups.

In addition to massive hemorrhage and subsequent multiple organ failure (MOF), colon necrosis is the most common cause of death in patients with rAAA.21 We also could not rescue one patient who died of broad colon necrosis during surgery. The complication rate of colon ischemia following rAAA surgery is 2–10%, and its mortality was estimated to be >50–80%.22 Reconstruction of the inferior mesenteric artery (IMA) is said to be the most effective technique in preventing colon necrosis, especially if the blood pressure of the IMA is <50 mmHg or if IMA/systolic blood pressure is <0.6.23, 24 The prevention of acute abdominal compartment syndrome is also important to improve outcome of rAAA. When primary abdominal closure without tension is difficult, secondary skin closure should be carried out to prevent circulatory or respiratory insufficiency.

Although EVAR under emergency circumstances improved survival of rAAA dramatically, as Mehta et al. reported,25 one randomized trial comparing EVAR and open repair for rAAA did not find a significant difference in mortality.26 Additional systematic data are pending on the results of EVAR for rAAA. From 2012, we also started using EVAR for selected rAAA patients, but there were still many problems to solve, like devising procurement or recovery from MOF secondary to acute compartment syndrome or hyperbilirubinemia induced by absorption of hematoma.

Limitations

As this is a retrospective study from a single center, logistic regression analysis for mortality was limited by the small number of events. Also, the relatively small number of patients might have resulted in insufficient statistical power. A large number of patients are needed to confirm these findings.

Conclusion

In our institute, 77.1% of patients who underwent emergency surgical treatment for all types of rAAA survived.

The value of BMI, serum lactate level, circulatory response following initial 2,000 mL volume loading, and volume of intraoperative blood loss were associated with the survival of patients who underwent surgical repair of rAAA.

Conflict of Interest

None.