- Email: [email protected]
Natural history and management of splanchnic artery aneurysms in a single tertiary referral center
From the Society for Vascular Surgery
Natural history and management of splanchnic artery aneurysms in a single tertiary referral center Young Erben, MD,a Adam J. Brownstein, BA,b Sareh Rajaee, MD,a Yupeng Li, PhD,c John A. Rizzo, PhD,b,c Hamid Mojibian, MD,d Bulat A. Ziganshin, MD, PhD,b,e and John A. Elefteriades, MD,b New Haven, Conn; Stony Brook, NY; and Kazan, Russia
ABSTRACT Objective: Splanchnic artery aneurysms (SAAs) are rare, and little is known about their natural history and management. We reviewed our single-center experience in managing this population of patients. Methods: A retrospective review of the Yale radiologic database from January 1999 to December 2016 was performed. Only patients with an SAA and a computed tomography scan of the abdomen were selected for review. Demographics of the patients, aneurysm characteristics, management, postoperative complications, and follow-up data were collected. Our primary outcomes included aneurysm growth rate and risk of rupture in those patients managed nonoperatively and morbidity and mortality of those SAA patients who underwent operative intervention. Results: There were 122 patients with 138 SAAs identified; 77 were male (62%), with a mean age of 66 years (range, 25-94 years). On computed tomography, 56 (45%) had previously diagnosed or concomitant aneurysms elsewhere. Of the patients managed nonoperatively, 101 patients (79%) had 108 SAAs; in the operative intervention group, 25 (21%) patients had 30 SAAs. The mean overall vessel diameter was 1.76 6 0.83 cm. The diameter of observed and operatively repaired SAAs was 1.58 6 0.56 cm and 2.41 6 1.23 cm, respectively (P ¼ .00001). Mean follow-up was 50 6 42 months for nonoperative management without any adverse events related to SAA, including 10 patients with SAA >2.0 cm. The mean observed growth rate for SAA was 0.064 6 0.18 cm/y. All symptomatic patients who presented with severe abdominal pain (n ¼ 11 [44%]) underwent operative intervention. Five patients presented with a ruptured SAA (3.6%; range, 2.3-5.0 cm); all of them except one underwent operative intervention. Other indications for repair included large size in seven, rapid growth in two, other open abdominal surgical procedures in two, multiple aneurysms in one, and desire to pursue fertility treatment in one. Operative repair included 14 (56%) endovascular embolizations and 11 (44%) open abdominal operations. After endovascular embolization, two patients underwent abdominal operation for hemorrhage and splenectomy. Open repairs included bypasses in six, splenectomy in two, resection in two, and plication in one. Two patients had postoperative acute kidney injury that resolved and one died of multisystem organ failure. One bypass occluded without sequelae. On multivariable regression analysis, female sex (P ¼ .02) was associated with faster growth rate, and a history of smoking (P ¼ .04) was associated with slower growth rate. Conclusions: It seems reasonable to observe asymptomatic patients with an SAA <2.0 cm because of the slow growth rate (0.064 6 0.18 cm/y) and benign behavior. When intervention is needed, both open and endovascular options should be considered. (J Vasc Surg 2018;-:1-9.)
From the Section of Vascular and Endovascular Surgery,a Aortic Institute at Yale-New Haven Hospital,b and Section of Vascular Interventional Radiology,d Yale School of Medicine, New Haven; the Department of Economics and Department of Family, Population and Preventive Medicine, Stony Brook University, Stony Brookc; and the Department of Surgical Diseases #2, Kazan State Medical University, Kazan.e A.J.B. was supported by the Leon Rosenberg, MD, Medical Student Research Fellowship and the Richard A. Moggio, MD, Student Research Fellowship. Author conflict of interest: none. Poster presentation at the 2017 Vascular Annual Meeting of the Society for Vascular Surgery, San Diego, Calif, May 31-June 3, 2017. Correspondence: John A. Elefteriades, MD, Aortic Institute at Yale-New Haven Hospital, Yale University School of Medicine, 789 Howard Ave, Clinic Bldg CB317, New Haven, CT 06519 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2018 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2017.12.057
Based on autopsy reports, the prevalence of splanchnic artery aneurysms (SAAs) may be as high as 10%.1 Splanchnic arteries include the celiac artery, superior mesenteric artery, and inferior mesenteric artery and all of its branches.2 It is important to recognize SAAs because up to 25% may be complicated by rupture, which has an estimated mortality rate between 25% and 70%. There are only a handful of series reporting SAAs, and most recommend early operative intervention without any natural history data to support their contention.3-7 Three large series to date report the importance of operative intervention in the presence of pseudoaneurysms and branch vessel aneurysms.2,8,9 Corey et al2 recently reported that 91% of small SAAs (<25 mm) remain stable in size and therefore may not require frequent surveillance imaging. Given the paucity of data on the management of SAAs, we sought to review our single-center experience with this population of patients. Our aim was to report the natural history of 1
2
Journal of Vascular Surgery
Erben et al
---
2018
Table I. Comorbidities and clinical presentation of the patients Total patients (N ¼ 120)
Patients undergoing operative intervention
Patients managed nonoperatively
P
66 6 14 (29-94)
58 6 16 (25-79)
68 6 12 (29-94)
<.00001
109 (89)
21 (84)
88 (91)
.465
77 (63)
14 (56)
63 (65)
.487
Asymptomatic
111 (91)
14 (56)
97 (100)
<.0001
Symptomatic
11 (9)
11 (44)
0 (0)
<.0001
Hypertension
79 (68)
17 (71)
62 (67)
.811
Smoking history
59 (56)
11 (50)
48 (57)
.632
Hyperlipidemia
56 (49)
8 (33)
48 (53)
.111
CAD
25 (22)
2 (9)
23 (25)
.096
CKD
19 (16)
1 (4)
18 (20)
.117
Diabetes mellitus
17 (15)
1 (4)
16 (18)
.118
Atrial fibrillation
16 (14)
4 (17)
12 (13)
.741
COPD
13 (11)
0 (0)
13 (14)
.067
History of stroke or TIA
9 (8)
1 (4)
8 (9)
.682
CHF
8 (7)
1 (4)
7 (8)
1
History of MI
7 (6)
1 (4)
6 (7)
1
PAD
5 (4)
1 (4)
4 (4)
1
Connective tissue disorder (Ehlers-Danlos syndrome)
1 (1)
0 (0)
1 (1)
1
Characteristic Demographics Age, years, mean 6 SD (range) White race Male Presentation
Comorbidities
CAD, Coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; MI, myocardial infarction; PAD, peripheral arterial disease; SD, standard deviation; TIA, transient ischemic attack. Values are reported as number (%) unless otherwise indicated.
SAAs and the outcomes of those SAAs treated in an open or endovascular fashion.
METHODS Cohort of patients. We identified all patients with SAAs at Yale-New Haven Hospital from January 1, 1999, to December 31, 2016. Patients were identified from the Yale radiology database through specific search terms and by International Classification of Diseases, Ninth Revision disease codes 442.84 (visceral aneurysm), 442.83 (splenic artery aneurysm), and 442.89 (aneurysm of other specified artery). Furthermore, only those patients with a computed tomography (CT) scan were included in the study. Data collection included details of each patient’s clinical presentation, comorbidities, management, and follow-up data as identified through retrospective chart review (Table I). Our primary outcomes included aneurysm growth rate and risk of rupture in those patients managed nonoperatively and morbidity and mortality of those SAA patients who underwent operative intervention. Consent of individual patients for study inclusion was not obtained or required, and this study received approval from the Yale Institutional Review Board (HIC# 1411014866) as a retrospective chart review.
Definitions. An aneurysm of the splanchnic arteries was defined as a lesion that is 1.5 times the size of the native vessel on axial imaging, reported by our radiologists. The decision for intervention was based on the individual surgeon’s preference. Patients were considered lost to follow-up if they did not undergo any additional CT imaging after the initial diagnostic cross-sectional study. Acute kidney injury was defined using the Kidney Disease: Improving Global Outcomes definition. This includes an increase in serum creatinine concentration by $0.3 mg/dL within 48 hours, an increase in serum creatinine concentration $1.5 times baseline, and a urine volume <0.5 mL/kg/h for 6 hours. Imaging data and growth rate calculation. CT scans were used to determine the size and location of the SAA. Aneurysm diameter was determined by directly measuring aneurysm size from outer wall to outer wall. Observed growth rate was calculated for all patients with two CT scans as follows: (diameter at last CT scan e diameter at initial CT scan)/time interval between CT scans. Also, a statistically estimated growth rate for SAA was calculated with an instrumental variables approach. The estimates were obtained by means of regression analysis in which aneurysm growth followed
Journal of Vascular Surgery Volume
-,
Number
Erben et al
3
-
Table II. Distribution of splanchnic artery aneurysms (SAAs) Location
No. (%)
Mean initial size, cm
Mean final size, cm
Celiac
63 (46)
1.46 6 0.31
1.53 6 0.34
Splenic
42 (30)
1.65 6 0.91
1.75 6 0.94
2 (0.6 and 2.3 cm)
SMA
12 (9)
2.50 6 1.30
2.70 6 1.20
1 (3.6 cma)
SMA branch
8 (6)
1.60 6 0.53
1.60 6 0.53
0
CHA
6 (4)
2.13 6 1.36
2.67 6 1.22
1 (5.0 cm)
PDA
3 (2)
2.0 6 0.75
2.07 6 0.66
0
GDA
1 (1)
2.3
2.3
1 (2.3 cm)
IMA
1 (1)
0.6
0.6
0
IMA branch
1 (1)
1.3
1.3
0
Left gastric
1 (1)
1.1
1.1
0
1.60 6 0.80
1.76 6 0.83
5
Total
138
No. of ruptures (size) 0 a
CHA, Common hepatic artery; GDA, gastroduodenal artery; IMA, inferior mesenteric artery; PDA, pancreaticoduodenal artery; SMA, superior mesenteric artery. a Pseudoaneurysms, which were not included in the growth rate calculation.
Celiac n=63 CHA (46%) n=6 GDA (4%) PDA n=3 (2%) SMA branch n=8 (6%)
n=1 (1%)
SMA n=12 (9%)
LGA n=1 (1%)
and to guard against disproportionately overweighing data on patients with stable aneurysms.
Splenic n=42 (30%)
IMA n=1 (1%) IMA branch n=1 (1%)
Fig 1. Distribution of all splanchnic artery aneurysms (SAAs) identified in 122 patients. CHA, Common hepatic artery; GDA, gastroduodenal artery; IMA, inferior mesenteric artery; LGA, left gastric artery; PDA, pancreaticoduodenal artery; SMA, superior mesenteric artery.
an exponential path. In particular, the natural logarithm of the last measured size to the first measured size was related to the time interval between the two tests and the disturbance term. Because the dependent variable has been transformed in the natural log form, we use Duan’s smearing estimator to calculate the expected value of the untransformed dependent variable. This method, which has been previously described by Rizzo et al10 in 1998, has been shown to be more accurate for measuring aneurysm growth. This method was employed to minimize the impact of measurement errors on SAA growth rate calculations
Statistical analysis. Our primary outcomes included aneurysm growth rate and risk of rupture in those patients managed nonoperatively and morbidity and mortality of those SAA patients who underwent operative intervention. Data are reported using means and standard deviations for continuous variables or frequencies for categorical variables. Student t-tests were used to analyze the differences in aneurysm size between operative intervention and nonoperative management groups. Fisher exact tests were used to compare comorbidity rates between nonoperative management and operative intervention groups. Multivariable regression analysis was used to evaluate the impact of comorbidities and aneurysm characteristics on growth rate. Statistical analyses were performed using R 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria) and vassarstats.net.
RESULTS Patients. We identified 122 patients with 138 SAAs. There were 77 male (63%) and 45 (37%) female patients. The mean age at presentation was 66 6 14 years (range, 25-94 years). The 25 patients in the operative intervention group were younger than the ones in the nonoperatively managed group (55 6 16 years vs 68 6 12 years; P < .00001). The majority of patients were asymptomatic at the time of the initial imaging study, and therefore they were observed without operative intervention (97 vs 14 patients in the nonoperatively managed and operative intervention groups, respectively; P < .0001). All patients’ comorbidities at presentation are shown in Table I. These included hypertension (68%), history of smoking (56%), hyperlipidemia (49%), coronary artery disease (22%), chronic kidney disease (16%), diabetes mellitus (15%), atrial fibrillation (14%), chronic obstructive
4
Journal of Vascular Surgery
Erben et al
---
2018
Fig 2. Axial cuts and three-dimensional reconstruction of computed tomography (CT) images demonstrating, at the arrow, (A and B) celiac artery aneurysm, (C) gastroduodenal artery aneurysm rupture, (D) superior mesenteric artery aneurysm rupture, (E) splenic artery aneurysm, and (F) inferior mesenteric artery aneurysm.
Table III. Breakdown of aneurysm locations for patients with multiple aneurysms Location of aneurysms
No. (%)
Ascending thoracic aorta
18 (15)
Renal artery
17 (14)
Other splanchnic arteries
15 (12)
Abdominal aorta
14 (11)
Iliac artery
11 (9)
Descending thoracic aorta
4 (3)
Thoracoabdominal aorta
3 (2)
Left subclavian artery
1 (1)
Internal carotid artery Total No. of patients
1 (1) 56 (46)
pulmonary disease (11%), history of stroke and transient ischemic attack (8%), congestive heart failure (7%), history of myocardial infarction (6%), peripheral arterial disease (4%), and connective tissue disorder (1%). The specific location of each of the SAAs is tabulated in Table II, schematically noted in Fig 1 and in cross-sectional imaging in Fig 2. The most common SAA identified was the celiac artery (46%), followed by the splenic artery (30%), the superior mesenteric artery (9%), a branch of the superior mesenteric artery (6%), the common hepatic artery (4%), the pancreaticoduodenal artery (2%), the gastroduodenal artery (1%), the inferior mesenteric artery (1%), a branch of the inferior mesenteric artery (1%), and the left gastric artery (1%). Eleven (9%) patients presented with abdominal pain, and five aneurysms (3.6%) were ruptured. The rest were asymptomatic. All symptomatic patients underwent
operative intervention. Of the five aneurysms that ruptured, two were pseudoaneurysms and three were true aneurysms. These pseudoaneurysms were not taken into consideration in the calculation of growth rate because their natural history is different from that of true aneurysms. The pseudoaneurysms were located in the superior mesenteric artery and splenic artery and were 3.6 cm and 0.6 cm in diameter, respectively. The causes of the pseudoaneurysms were acute or chronic pancreatitis and embolic of an infectious nature in a patient with endocarditis and multiple other infections including hepatitis C and human immunodeficiency virus infection. The true aneurysms that ruptured were located in the common hepatic, gastroduodenal, and splenic arteries and were 5.0 cm, 2.3 cm, and 2.3 cm in diameter, respectively. Concomitant aneurysms elsewhere in the vascular tree were identified in 56 (46%) patients (Table III). The locations included the ascending thoracic aorta (15%), renal artery (14%), other splanchnic arteries (12%), abdominal aorta (11%), iliac artery (9%), descending thoracic aorta (3%), thoracoabdominal aorta (2%), left subclavian artery (1%), and internal carotid artery (1%). Observed SAAs (nonoperatively managed group). The management of all SAAs is shown in Fig 3. There were 101 patients with 108 SAAs who were observed without operative intervention. Seventy-five patients had two CT scans that were on average 50 6 42 months apart. Twenty-six patients with 28 SAAs were lost to follow-up. SAAs were followed up for 50 6 42 months and demonstrated a slow observed mean growth rate of 0.064 6 0.18 cm/y without any ruptures (Fig 4, A). Of note,
Journal of Vascular Surgery Volume
-,
Number
Erben et al
5
-
Fig 3. Management of patients with splanchnic artery aneurysms (SAAs). NOR, Nonoperative management; OR, operative intervention.
Fig 4. A, Frequency distribution of observed growth rates. B, Frequency distribution of estimated growth rates.
10 patients had an SAA >2.0 cm in diameter. The frequency distribution of estimated growth rates is demonstrated in Fig 4, B and was calculated to be 0.086 6 0.029 cm/y. To evaluate the impact of comorbidities on growth rates, a multivariable regression analysis was conducted (Table IV). Only female sex and smoking history were found to be statistically significant. Female sex (P ¼ .02; odds ratio, 1.020; 95% confidence interval, 1.003-1.037) was associated with a faster growth rate; a history of smoking (P ¼ .04; odds ratio, 0.981; 95% confidence interval, 0.963-0.999) was associated with a slower growth rate. All other variables evaluated including age, chronic obstructive pulmonary disease, coronary artery disease, connective tissue disorder, diabetes mellitus, chronic kidney disease, history of stroke or transient ischemic attack, hypertension, congestive heart failure, peripheral arterial disease, and presence of multiple aneurysms did not contribute to the growth of SAAs.
SAAs treated (operative intervention group). Twentyfive patients with 30 SAAs underwent operative repair (Fig 2). The mean aneurysm size was 2.41 6 1.23 cm (the mean size of SAAs in the nonoperatively managed group was 1.58 6 0.56 cm; P ¼ .00001). Four of the five patients who presented with ruptured SAAs underwent repair. The 0.6-cm ruptured splenic artery pseudoaneurysm underwent coil embolization; the 3.6-cm ruptured superior mesenteric artery pseudoaneurysm underwent open bypass. For the true aneurysms that ruptured, the 2.3-cm gastroduodenal artery aneurysm underwent coil embolization, whereas the 2.3-cm splenic artery aneurysm underwent ligation with splenectomy. The patient with the 5.0-cm ruptured common hepatic artery aneurysm chose not to undergo intervention and died. All other nonruptured aneurysms underwent repair because of large size (23%), rapid growth (7%), other open abdominal surgical procedure required (7%), multiple aneurysms (3%), and desire to pursue fertility
6
Journal of Vascular Surgery
Erben et al
---
2018
Table IV. Multivariable regression analysis of the impact of comorbidities on growth rate Variable
P value (odds ratio, 95% confidence interval)
Direction of impact on SAA growth
Female sex
Increased
Smoking history
Decreased
.024 (1.02, 1.003-1.037) .041 (0.98, 0.963-0.999)
Age
d
.097 (1.00, 0.999-1.00)
COPD
d
.086 (1.02, 0.996-1.054)
CAD
d
.605 (0.99, 0.975-1.015)
Connective tissue disease
d
.903 (0.98, 0.738-1.308)
Diabetes mellitus
d
.459 (0.99, 0.975-1.012)
CKD
d
.542 (1.01, 0.988-1.023)
History of stroke or TIA
d
.730 (0.993, 0.956-1.032)
Hypertension
d
.569 (1.005, 0.988-1.023) .940 (1.00, 0.946-1.062)
CHF
d
PAD
d
.841 (0.99, 0.881-1.109)
Multiple aneurysms
d
.068 (1.01, 0.999-1.029)
CAD, Coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; PAD, peripheral arterial disease; SAA, splanchnic artery aneurysm; TIA, transient ischemic attack.
Table V. Location and mode of repair for splanchnic artery aneurysms (SAAs) Location
No.
Size, cm
Endovascular repair, No. (%)
Open repair, No. (%)
Type of open repair (No.)
12
2.44 6 1.31
9 (75)
3 (25)
Splenectomy (2), resection (1)
SMA
5
3.38 6 1.46
1 (20)
4 (80)
Bypass (4)
SMA branch
3
1.88 6 0.54
3 (75)
1 (25)
Bypass (1)
CA
3
1.43 6 0.26
2 (67)
1 (33)
Bypass (1)
CHA
2
2.95 6 0.35
2 (100)
PDA
2
2.50 6 0.30
1 (50)
GDA
1
2.3
IMA
1
0.6
Splenic
1 (100) 0 (0)
0 (0) 1 (50) 0 (0) 1 (100)
d Plication (1) d Resection (1)
CA, Celiac artery; CHA, common hepatic artery; GDA, gastroduodenal artery; IMA, inferior mesenteric artery; PDA, pancreaticoduodenal artery; SMA, superior mesenteric artery.
treatment (3%) at a mean time of 35.4 6 23.5 months from initial presentation. The repair modality is illustrated in Table V and included both endovascular coil embolization and open surgical techniques, including bypass (55%), splenectomy (18%), resection (18%), and plication (9%). Complications (Table VI) related to the procedure or operation occurred in six patients (20%), which included two in the coil embolization group: one open exploration due to hemorrhage and one splenectomy. Postoperative complications from the open surgical group included acute kidney injury in two patients that resolved without any intervention and death due to multisystem organ failure in one patient after open bypass repair of a 4.6-cm superior mesenteric artery aneurysm. Last, one superior mesenteric artery bypass was found to be occluded on follow-up, incidentally detected on crosssectional imaging, without any sequelae.
DISCUSSION Our series is the first and only one with longitudinal follow-up and growth rate calculations in patients
with SAAs. We have been able to determine the growth rate of SAAs at 0.064 6 0.18 cm/y while following their natural history. Furthermore, we have analyzed the outcomes of those SAAs that underwent operative repair, including both endovascular and open techniques. There are few data available in the literature regarding the natural history of SAAs. Most available series have a small number of patients and recommend aggressive intervention at a diameter >2.0 cm because of an anticipated high mortality rate.11,12 One aspect that remains constant in the reviews available is the lethality of SAAs in pregnant women. Studies report a maternal mortality rate of 75% and fetal mortality rate of 95% for ruptured SAA in pregnancy, thereby explaining the emphasis on early intervention.13-15 We report in our cohort that the desire to pursue fertility treatment in the presence of an SAA was an indication to undergo repair of one aneurysm, which is certainly in agreement with the reported literature.16-20
Journal of Vascular Surgery Volume
-,
Number
Erben et al
7
-
Table VI. Complications after open operative repair Outcome
No. (%)
Details
Mortality
1 (3)
Multisystem organ failure
Overall complications
6 (20)
Hemorrhage
1 (3)
Open exploration
Splenectomy
1 (3)
After embolization
Acute kidney injury
2 (7)
After embolization
Bypass occlusion
1 (3)
Time unknown
Aneurysm location. One important aspect that we encountered in our cohort was the relatively large number of celiac artery aneurysms (46% instead of <5%) and the relatively low number of splenic artery aneurysms (30% instead of >60%) that are mostly reported in the literature.12,21 We believe the reason for this selection bias is that we used cross-sectional imaging as the only imaging technique in our study to ensure that all patients were measured by the same objective imaging modality to allow fair comparison on follow-up. The majority of the series reporting splenic artery aneurysms included ultrasound as a major diagnostic tool, which could have biased our selection of patients.22-25 Furthermore, once splenic artery aneurysms are diagnosed, they undergo angiography for endovascular repair and skip the cross-sectional imaging; this negatively affects our selection of patients in using a cross-sectional imaging database search.26 Furthermore, our study includes a retrospective review of radiologic data and medical records, which could as well have influenced the selection bias because we depend on the medical record coding and dictation by the radiologist reading the cross-sectional imaging study. Observed SAAs (nonoperatively managed group). A most recent review by Corey et al2 recommended SAA surveillance and aggressive intervention in aneurysms of the pancreaticoduodenal arcade and gastroduodenal artery because these were most likely to rupture in their cohort. In our series, we were able to demonstrate a slow growth rate in SAAs of 0.064 6 0.18 cm/y and that the smallest true aneurysm rupture occurred at 2.3 cm in diameter in small-caliber arteries (splenic and gastroduodenal arteries). Unfortunately, in our series, we encountered only a handful of pancreaticoduodenal and gastroduodenal artery aneurysms (n ¼ 4), which hinders the direct correlation of these smaller caliber vessels and their higher rupture risk. Although the smallest rupture occurred at 2.3 cm, in our cohort, we observed 10 patients with SAAs >2.0 cm without any rupture. Therefore, it seems feasible to observe all SAAs with a very low risk of rupture. On multivariable regression analysis, female sex was associated with a faster growth rate, whereas a history of smoking was associated with a slower growth rate of
SAAs. Therefore, women may have a higher risk of adverse outcomes in the face of SAA, which has been reported in multiple prior studies, justifying a higher risk of mortality in the female patient with SAA.27,28 Another important factor is the known correlation between estrogen and aortic medial degeneration, which could potentially explain the risk of female patients for a worse outcome and relative protection of male patients with SAA.29,30 Interestingly, it has been shown that smoking is a risk factor for growth in the face of both an infrarenal abdominal aortic aneurysm31 and an intracranial aneurysm; however, the exact mechanism is not well understood.32,33 Our analysis demonstrated that smoking was associated with slower growth of SAAs. However, we are not sure how to interpret these results. It may be possible that these results do not apply in the growth rate of SAA. These data should be interpreted with caution. With the current data available from this study, we would recommend in patients with small SAAs yearly cross-sectional imaging follow-up because of the slow rate of growth in these SAAs. These findings, although encouraging, represent the first natural history study with growth rate calculations. Therefore, we would welcome and encourage other natural history projects to corroborate these data. SAAs treated (operative intervention group). In our review, we identified 30 SAAs in 25 patients who underwent operative repair (Fig 3). Six of these presented ruptured with two pseudoaneurysms of the splenic and superior mesenteric artery measuring 0.6 cm and 3.6 cm in diameter. There is general consensus to repair pseudoaneurysms because of their unpredictable risk of rupture at any size,34 hence their immediate repair on presentation in our cohort. All true aneurysms that underwent operative intervention had a mean size of 2.41 6 1.23 cm vs 1.58 6 0.56 cm in those patients whose SAAs were observed (P ¼ .00001). Our operative morbidity and mortality (20% and 3%, respectively) were comparable to the ones reported in the literature.2,9,11 One patient died of multisystem organ failure after a bypass for a superior mesenteric artery aneurysm of 4.6 cm in diameter. Furthermore, the operative explorations from those patients treated with endovascular techniques were expected complications. One patient underwent exploration because of hemorrhage from a ruptured gastroduodenal artery with hemodynamic instability, for whom no bleeding source was found, and a second patient with a necrotic spleen required splenectomy. The complications in those patients who underwent open repair included acute kidney injury, which resolved without any intervention, and an occluded bypass at an unknown time after operation without any negative sequelae.
8
Journal of Vascular Surgery
Erben et al
---
We believe that a strength of our natural history report using the radiologic database for SAAs is that it offers a true representation of all diagnosed patients with this disease entity. Most are asymptomatic, small aneurysms with low risk of rupture. Previous reports are derived from surgical databases, and therefore most of these patients are already preselected for operative intervention.2,6,7 By observing all patients with abnormalities on CT scans, we are able to assess the natural history of SAAs, including incidentally found SAAs. Study limitations. We have encountered several limitations with the current project, which include the relatively small number of aneurysms analyzed and the retrospective nature of our review. Furthermore, to objectively assess growth rates, we used cross-sectional imaging in the form of CT to assess the size of the vessels in question. Therefore, a large proportion of patients diagnosed by other imaging modalities have not been taken into consideration, which could have clearly strengthened the study by the sheer number of patients. However, because of unpredictability of ultrasound measurements, these were not considered. Next, if codes for SAA diagnosis were not entered properly within the electronic medical records of patients, these potential patients with SAA were not picked up at the time of data collection.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
CONCLUSIONS Our report regarding the natural history of SAAs is able to calculate growth rates for SAAs. The growth rate of SAAs is very slow (0.064 6 0.18 cm/y), and therefore it seems reasonable to observe asymptomatic patients with a small to medium-sized SAA. In our series, the smallest ruptured SAA occurred at 2.3 cm in diameter. If operative intervention should be needed, both open and endovascular options can be considered.
13.
14.
15.
AUTHOR CONTRIBUTIONS Conception and design: YE, AB, JE Analysis and interpretation: YE, AB, SR, YL, JR, HM, BZ, JE Data collection: YE, AB, SR, HM Writing the article: YE, AB, BZ, JE Critical revision of the article: YE, AB, SR, YL, JR, HM, BZ, JE Final approval of the article: YE, AB, SR, YL, JR, HM, BZ, JE Statistical analysis: AB, YL, JR Obtained funding: Not applicable Overall responsibility: JE YE and AB contributed equally to this article and share first authorship.
16.
17.
18.
19.
20.
REFERENCES 1. Bedford PD, Lodge B. Aneurysm of the splenic artery. Gut 1960;1:312-20. 2. Corey MR, Ergul EA, Cambria RP, English SJ, Patel VI, Lancaster RT, et al. The natural history of splanchnic artery
21. 22.
2018
aneurysms and outcomes after operative intervention. J Vasc Surg 2016;63:949-57. Abbas MA, Fowl RJ, Stone WM, Panneton JM, Oldenburg WA, Bower TC, et al. Hepatic artery aneurysm: factors that predict complications. J Vasc Surg 2003;38: 41-5. Abbas MA, Stone WM, Fowl RJ, Gloviczki P, Oldenburg WA, Pairolero PC, et al. Splenic artery aneurysms: two decades experience at Mayo clinic. Ann Vasc Surg 2002;16:442-9. Saltzberg SS, Maldonado TS, Lamparello PJ, Cayne NS, Nalbandian MM, Rosen RJ, et al. Is endovascular therapy the preferred treatment for all visceral artery aneurysms? Ann Vasc Surg 2005;19:507-15. Stone WM, Abbas M, Cherry KJ, Fowl RJ, Gloviczki P. Superior mesenteric artery aneurysms: is presence an indication for intervention? J Vasc Surg 2002;36:234-7; discussion: 237. Stone WM, Abbas MA, Gloviczki P, Fowl RJ, Cherry KJ. Celiac arterial aneurysms: a critical reappraisal of a rare entity. Arch Surg 2002;137:670-4. Tetreau R, Beji H, Henry L, Valette PJ, Pilleul F. Arterial splanchnic aneurysms: presentation, treatment and outcome in 112 patients. Diagn Interv Imaging 2016;97: 81-90. Shukla AJ, Eid R, Fish L, Avgerinos E, Marone L, Makaroun M, et al. Contemporary outcomes of intact and ruptured visceral artery aneurysms. J Vasc Surg 2015;61:1442-7. Rizzo JA, Coady MA, Elefteriades JA. Procedures for estimating growth rates in thoracic aortic aneurysms. J Clin Epidemiol 1998;51:747-54. Panayiotopoulos YP, Assadourian R, Taylor PR. Aneurysms of the visceral and renal arteries. Ann R Coll Surg Engl 1996;78: 412-9. van Rijn MJ, Ten Raa S, Hendriks JM, Verhagen HJ. Visceral aneurysms: old paradigms, new insights? Best Pract Res Clin Gastroenterol 2017;31:97-104. Hunsaker DM, Turner S, Hunsaker JC 3rd. Sudden and unexpected death resulting from splenic artery aneurysm rupture: two case reports of pregnancy-related fatal rupture of splenic artery aneurysm. Am J Forensic Med Pathol 2002;23:338-41. Veluppillai C, Perreve S, de Kerviler B, Ducarme G. [Splenic arterial aneurysm and pregnancy: a review]. Presse Med 2015;44:991-4. Stanley JC, Wakefield TW, Graham LM, Whitehouse WM Jr, Zelenock GB, Lindenauer SM. Clinical importance and management of splanchnic artery aneurysms. J Vasc Surg 1986;3:836-40. Benali M, Charrada H, Bouassida M, Bahloul A, Jmal K, Dhouib F, et al. [Splenic artery aneurysm rupture in late pregnancy: a case report]. Ann Fr Anesth Reanim 2013;32: 721-2. Lang W, Strobel D, Beinder E, Raab M. Surgery of a splenic artery aneurysm during pregnancy. Eur J Obstet Gynecol Reprod Biol 2002;102:215-6. Shahabi S, Jani J, Masters L, Cobin L, Greindl J. Spontaneous rupture of a splenic artery aneurysm in pregnancy: report of two cases. Acta Chir Belg 2000;100:231-3. Arrieta FM, Muguerza J, Garcia Sancho L, Ayuso M, Rustarazu M, Valenzuela P. Rupture of splenic artery aneurysm during pregnancy and posterior evolution of gestation. Zentralbl Gynakol 2000;122:579-80. Cressey D, Reid MF. Splenic artery aneurysm rupture in pregnancy. Int J Obstet Anesth 1996;5:103-4. Pasha SF, Gloviczki P, Stanson AW, Kamath PS. Splanchnic artery aneurysms. Mayo Clin Proc 2007;82:472-9. Fransen H, Kubale R, Wurche KD, Kalahne A. [Noninvasive diagnosis of splenic artery aneurysm]. Rofo 1989;151:532-5.
Journal of Vascular Surgery Volume
-,
Number
Erben et al
9
-
23. Davis T, Minardi J, Knight J, Larrabee H, Schaefer G. Ruptured splenic artery aneurysm: rare cause of shock diagnosed with bedside ultrasound. West J Emerg Med 2015;16:762-5. 24. Lo WL, Mok KL. Ruptured splenic artery aneurysm detected by emergency ultrasoundda case report. Crit Ultrasound J 2015;7:26. 25. Dave SP, Reis ED, Hossain A, Taub PJ, Kerstein MD, Hollier LH. Splenic artery aneurysm in the 1990s. Ann Vasc Surg 2000;14:223-9. 26. Hogendoorn W, Lavida A, Hunink MG, Moll FL, Geroulakos G, Muhs BE, et al. Open repair, endovascular repair, and conservative management of true splenic artery aneurysms. J Vasc Surg 2014;60:1667-76.e1. 27. Ha JF, Phillips M, Faulkner K. Splenic artery aneurysm rupture in pregnancy. Eur J Obstet Gynecol Reprod Biol 2009;146:133-7. 28. Lynch MJ, Woodford NW. Rupture of a splenic artery aneurysm during pregnancy with maternal and foetal death: a case report. Med Sci Law 2008;48:342-5. 29. Crawford JD, Hsieh CM, Schenning RC, Slater MS, Landry GJ, Moneta GL, et al. Genetics, pregnancy, and aortic degeneration. Ann Vasc Surg 2016;30:158.e5-9.
30. Tripathi R, Sainathan S, Ziganshin BA, Elefteriades JA. Thoracic aortic aneurysm from chronic antiestrogen therapy. Int J Angiol 2017;26:60-3. 31. Enriquez-Vega ME, Solorio-Rosete HF, Cossio-Zazueta A, Bizueto-Rosas H, Cruz-Castillo JE, Iturburu-Enriquez A. [Early detection of abdominal aortic aneurysm in risk population]. Rev Med Inst Mex Seguro Soc 2015;53(Suppl 1):S100-3. 32. Backes D, Rinkel GJ, Greving JP, Velthuis BK, Murayama Y, Takao H, et al. ELAPSS score for prediction of risk of growth of unruptured intracranial aneurysms. Neurology 2017;88: 1600-6. 33. Kim HW, Stansfield BK. Genetic and epigenetic regulation of aortic aneurysms. Biomed Res Int 2017;2017:7268521. 34. Zyromski NJ, Vieira C, Stecker M, Nakeeb A, Pitt HA, Lillemoe KD, et al. Improved outcomes in postoperative and pancreatitis-related visceral pseudoaneurysms. J Gastrointest Surg 2007;11:50-5.
Submitted Oct 17, 2017; accepted Dec 30, 2017.
Natural history and management of splanchnic artery aneurysms in a single tertiary referral center Young Erben, MD,a Adam J. Brownstein, BA,b Sareh Rajaee, MD,a Yupeng Li, PhD,c John A. Rizzo, PhD,b,c Hamid Mojibian, MD,d Bulat A. Ziganshin, MD, PhD,b,e and John A. Elefteriades, MD,b New Haven, Conn; Stony Brook, NY; and Kazan, Russia
ABSTRACT Objective: Splanchnic artery aneurysms (SAAs) are rare, and little is known about their natural history and management. We reviewed our single-center experience in managing this population of patients. Methods: A retrospective review of the Yale radiologic database from January 1999 to December 2016 was performed. Only patients with an SAA and a computed tomography scan of the abdomen were selected for review. Demographics of the patients, aneurysm characteristics, management, postoperative complications, and follow-up data were collected. Our primary outcomes included aneurysm growth rate and risk of rupture in those patients managed nonoperatively and morbidity and mortality of those SAA patients who underwent operative intervention. Results: There were 122 patients with 138 SAAs identified; 77 were male (62%), with a mean age of 66 years (range, 25-94 years). On computed tomography, 56 (45%) had previously diagnosed or concomitant aneurysms elsewhere. Of the patients managed nonoperatively, 101 patients (79%) had 108 SAAs; in the operative intervention group, 25 (21%) patients had 30 SAAs. The mean overall vessel diameter was 1.76 6 0.83 cm. The diameter of observed and operatively repaired SAAs was 1.58 6 0.56 cm and 2.41 6 1.23 cm, respectively (P ¼ .00001). Mean follow-up was 50 6 42 months for nonoperative management without any adverse events related to SAA, including 10 patients with SAA >2.0 cm. The mean observed growth rate for SAA was 0.064 6 0.18 cm/y. All symptomatic patients who presented with severe abdominal pain (n ¼ 11 [44%]) underwent operative intervention. Five patients presented with a ruptured SAA (3.6%; range, 2.3-5.0 cm); all of them except one underwent operative intervention. Other indications for repair included large size in seven, rapid growth in two, other open abdominal surgical procedures in two, multiple aneurysms in one, and desire to pursue fertility treatment in one. Operative repair included 14 (56%) endovascular embolizations and 11 (44%) open abdominal operations. After endovascular embolization, two patients underwent abdominal operation for hemorrhage and splenectomy. Open repairs included bypasses in six, splenectomy in two, resection in two, and plication in one. Two patients had postoperative acute kidney injury that resolved and one died of multisystem organ failure. One bypass occluded without sequelae. On multivariable regression analysis, female sex (P ¼ .02) was associated with faster growth rate, and a history of smoking (P ¼ .04) was associated with slower growth rate. Conclusions: It seems reasonable to observe asymptomatic patients with an SAA <2.0 cm because of the slow growth rate (0.064 6 0.18 cm/y) and benign behavior. When intervention is needed, both open and endovascular options should be considered. (J Vasc Surg 2018;-:1-9.)
From the Section of Vascular and Endovascular Surgery,a Aortic Institute at Yale-New Haven Hospital,b and Section of Vascular Interventional Radiology,d Yale School of Medicine, New Haven; the Department of Economics and Department of Family, Population and Preventive Medicine, Stony Brook University, Stony Brookc; and the Department of Surgical Diseases #2, Kazan State Medical University, Kazan.e A.J.B. was supported by the Leon Rosenberg, MD, Medical Student Research Fellowship and the Richard A. Moggio, MD, Student Research Fellowship. Author conflict of interest: none. Poster presentation at the 2017 Vascular Annual Meeting of the Society for Vascular Surgery, San Diego, Calif, May 31-June 3, 2017. Correspondence: John A. Elefteriades, MD, Aortic Institute at Yale-New Haven Hospital, Yale University School of Medicine, 789 Howard Ave, Clinic Bldg CB317, New Haven, CT 06519 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2018 by the Society for Vascular Surgery. Published by Elsevier Inc. https://doi.org/10.1016/j.jvs.2017.12.057
Based on autopsy reports, the prevalence of splanchnic artery aneurysms (SAAs) may be as high as 10%.1 Splanchnic arteries include the celiac artery, superior mesenteric artery, and inferior mesenteric artery and all of its branches.2 It is important to recognize SAAs because up to 25% may be complicated by rupture, which has an estimated mortality rate between 25% and 70%. There are only a handful of series reporting SAAs, and most recommend early operative intervention without any natural history data to support their contention.3-7 Three large series to date report the importance of operative intervention in the presence of pseudoaneurysms and branch vessel aneurysms.2,8,9 Corey et al2 recently reported that 91% of small SAAs (<25 mm) remain stable in size and therefore may not require frequent surveillance imaging. Given the paucity of data on the management of SAAs, we sought to review our single-center experience with this population of patients. Our aim was to report the natural history of 1
2
Journal of Vascular Surgery
Erben et al
---
2018
Table I. Comorbidities and clinical presentation of the patients Total patients (N ¼ 120)
Patients undergoing operative intervention
Patients managed nonoperatively
P
66 6 14 (29-94)
58 6 16 (25-79)
68 6 12 (29-94)
<.00001
109 (89)
21 (84)
88 (91)
.465
77 (63)
14 (56)
63 (65)
.487
Asymptomatic
111 (91)
14 (56)
97 (100)
<.0001
Symptomatic
11 (9)
11 (44)
0 (0)
<.0001
Hypertension
79 (68)
17 (71)
62 (67)
.811
Smoking history
59 (56)
11 (50)
48 (57)
.632
Hyperlipidemia
56 (49)
8 (33)
48 (53)
.111
CAD
25 (22)
2 (9)
23 (25)
.096
CKD
19 (16)
1 (4)
18 (20)
.117
Diabetes mellitus
17 (15)
1 (4)
16 (18)
.118
Atrial fibrillation
16 (14)
4 (17)
12 (13)
.741
COPD
13 (11)
0 (0)
13 (14)
.067
History of stroke or TIA
9 (8)
1 (4)
8 (9)
.682
CHF
8 (7)
1 (4)
7 (8)
1
History of MI
7 (6)
1 (4)
6 (7)
1
PAD
5 (4)
1 (4)
4 (4)
1
Connective tissue disorder (Ehlers-Danlos syndrome)
1 (1)
0 (0)
1 (1)
1
Characteristic Demographics Age, years, mean 6 SD (range) White race Male Presentation
Comorbidities
CAD, Coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; MI, myocardial infarction; PAD, peripheral arterial disease; SD, standard deviation; TIA, transient ischemic attack. Values are reported as number (%) unless otherwise indicated.
SAAs and the outcomes of those SAAs treated in an open or endovascular fashion.
METHODS Cohort of patients. We identified all patients with SAAs at Yale-New Haven Hospital from January 1, 1999, to December 31, 2016. Patients were identified from the Yale radiology database through specific search terms and by International Classification of Diseases, Ninth Revision disease codes 442.84 (visceral aneurysm), 442.83 (splenic artery aneurysm), and 442.89 (aneurysm of other specified artery). Furthermore, only those patients with a computed tomography (CT) scan were included in the study. Data collection included details of each patient’s clinical presentation, comorbidities, management, and follow-up data as identified through retrospective chart review (Table I). Our primary outcomes included aneurysm growth rate and risk of rupture in those patients managed nonoperatively and morbidity and mortality of those SAA patients who underwent operative intervention. Consent of individual patients for study inclusion was not obtained or required, and this study received approval from the Yale Institutional Review Board (HIC# 1411014866) as a retrospective chart review.
Definitions. An aneurysm of the splanchnic arteries was defined as a lesion that is 1.5 times the size of the native vessel on axial imaging, reported by our radiologists. The decision for intervention was based on the individual surgeon’s preference. Patients were considered lost to follow-up if they did not undergo any additional CT imaging after the initial diagnostic cross-sectional study. Acute kidney injury was defined using the Kidney Disease: Improving Global Outcomes definition. This includes an increase in serum creatinine concentration by $0.3 mg/dL within 48 hours, an increase in serum creatinine concentration $1.5 times baseline, and a urine volume <0.5 mL/kg/h for 6 hours. Imaging data and growth rate calculation. CT scans were used to determine the size and location of the SAA. Aneurysm diameter was determined by directly measuring aneurysm size from outer wall to outer wall. Observed growth rate was calculated for all patients with two CT scans as follows: (diameter at last CT scan e diameter at initial CT scan)/time interval between CT scans. Also, a statistically estimated growth rate for SAA was calculated with an instrumental variables approach. The estimates were obtained by means of regression analysis in which aneurysm growth followed
Journal of Vascular Surgery Volume
-,
Number
Erben et al
3
-
Table II. Distribution of splanchnic artery aneurysms (SAAs) Location
No. (%)
Mean initial size, cm
Mean final size, cm
Celiac
63 (46)
1.46 6 0.31
1.53 6 0.34
Splenic
42 (30)
1.65 6 0.91
1.75 6 0.94
2 (0.6 and 2.3 cm)
SMA
12 (9)
2.50 6 1.30
2.70 6 1.20
1 (3.6 cma)
SMA branch
8 (6)
1.60 6 0.53
1.60 6 0.53
0
CHA
6 (4)
2.13 6 1.36
2.67 6 1.22
1 (5.0 cm)
PDA
3 (2)
2.0 6 0.75
2.07 6 0.66
0
GDA
1 (1)
2.3
2.3
1 (2.3 cm)
IMA
1 (1)
0.6
0.6
0
IMA branch
1 (1)
1.3
1.3
0
Left gastric
1 (1)
1.1
1.1
0
1.60 6 0.80
1.76 6 0.83
5
Total
138
No. of ruptures (size) 0 a
CHA, Common hepatic artery; GDA, gastroduodenal artery; IMA, inferior mesenteric artery; PDA, pancreaticoduodenal artery; SMA, superior mesenteric artery. a Pseudoaneurysms, which were not included in the growth rate calculation.
Celiac n=63 CHA (46%) n=6 GDA (4%) PDA n=3 (2%) SMA branch n=8 (6%)
n=1 (1%)
SMA n=12 (9%)
LGA n=1 (1%)
and to guard against disproportionately overweighing data on patients with stable aneurysms.
Splenic n=42 (30%)
IMA n=1 (1%) IMA branch n=1 (1%)
Fig 1. Distribution of all splanchnic artery aneurysms (SAAs) identified in 122 patients. CHA, Common hepatic artery; GDA, gastroduodenal artery; IMA, inferior mesenteric artery; LGA, left gastric artery; PDA, pancreaticoduodenal artery; SMA, superior mesenteric artery.
an exponential path. In particular, the natural logarithm of the last measured size to the first measured size was related to the time interval between the two tests and the disturbance term. Because the dependent variable has been transformed in the natural log form, we use Duan’s smearing estimator to calculate the expected value of the untransformed dependent variable. This method, which has been previously described by Rizzo et al10 in 1998, has been shown to be more accurate for measuring aneurysm growth. This method was employed to minimize the impact of measurement errors on SAA growth rate calculations
Statistical analysis. Our primary outcomes included aneurysm growth rate and risk of rupture in those patients managed nonoperatively and morbidity and mortality of those SAA patients who underwent operative intervention. Data are reported using means and standard deviations for continuous variables or frequencies for categorical variables. Student t-tests were used to analyze the differences in aneurysm size between operative intervention and nonoperative management groups. Fisher exact tests were used to compare comorbidity rates between nonoperative management and operative intervention groups. Multivariable regression analysis was used to evaluate the impact of comorbidities and aneurysm characteristics on growth rate. Statistical analyses were performed using R 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria) and vassarstats.net.
RESULTS Patients. We identified 122 patients with 138 SAAs. There were 77 male (63%) and 45 (37%) female patients. The mean age at presentation was 66 6 14 years (range, 25-94 years). The 25 patients in the operative intervention group were younger than the ones in the nonoperatively managed group (55 6 16 years vs 68 6 12 years; P < .00001). The majority of patients were asymptomatic at the time of the initial imaging study, and therefore they were observed without operative intervention (97 vs 14 patients in the nonoperatively managed and operative intervention groups, respectively; P < .0001). All patients’ comorbidities at presentation are shown in Table I. These included hypertension (68%), history of smoking (56%), hyperlipidemia (49%), coronary artery disease (22%), chronic kidney disease (16%), diabetes mellitus (15%), atrial fibrillation (14%), chronic obstructive
4
Journal of Vascular Surgery
Erben et al
---
2018
Fig 2. Axial cuts and three-dimensional reconstruction of computed tomography (CT) images demonstrating, at the arrow, (A and B) celiac artery aneurysm, (C) gastroduodenal artery aneurysm rupture, (D) superior mesenteric artery aneurysm rupture, (E) splenic artery aneurysm, and (F) inferior mesenteric artery aneurysm.
Table III. Breakdown of aneurysm locations for patients with multiple aneurysms Location of aneurysms
No. (%)
Ascending thoracic aorta
18 (15)
Renal artery
17 (14)
Other splanchnic arteries
15 (12)
Abdominal aorta
14 (11)
Iliac artery
11 (9)
Descending thoracic aorta
4 (3)
Thoracoabdominal aorta
3 (2)
Left subclavian artery
1 (1)
Internal carotid artery Total No. of patients
1 (1) 56 (46)
pulmonary disease (11%), history of stroke and transient ischemic attack (8%), congestive heart failure (7%), history of myocardial infarction (6%), peripheral arterial disease (4%), and connective tissue disorder (1%). The specific location of each of the SAAs is tabulated in Table II, schematically noted in Fig 1 and in cross-sectional imaging in Fig 2. The most common SAA identified was the celiac artery (46%), followed by the splenic artery (30%), the superior mesenteric artery (9%), a branch of the superior mesenteric artery (6%), the common hepatic artery (4%), the pancreaticoduodenal artery (2%), the gastroduodenal artery (1%), the inferior mesenteric artery (1%), a branch of the inferior mesenteric artery (1%), and the left gastric artery (1%). Eleven (9%) patients presented with abdominal pain, and five aneurysms (3.6%) were ruptured. The rest were asymptomatic. All symptomatic patients underwent
operative intervention. Of the five aneurysms that ruptured, two were pseudoaneurysms and three were true aneurysms. These pseudoaneurysms were not taken into consideration in the calculation of growth rate because their natural history is different from that of true aneurysms. The pseudoaneurysms were located in the superior mesenteric artery and splenic artery and were 3.6 cm and 0.6 cm in diameter, respectively. The causes of the pseudoaneurysms were acute or chronic pancreatitis and embolic of an infectious nature in a patient with endocarditis and multiple other infections including hepatitis C and human immunodeficiency virus infection. The true aneurysms that ruptured were located in the common hepatic, gastroduodenal, and splenic arteries and were 5.0 cm, 2.3 cm, and 2.3 cm in diameter, respectively. Concomitant aneurysms elsewhere in the vascular tree were identified in 56 (46%) patients (Table III). The locations included the ascending thoracic aorta (15%), renal artery (14%), other splanchnic arteries (12%), abdominal aorta (11%), iliac artery (9%), descending thoracic aorta (3%), thoracoabdominal aorta (2%), left subclavian artery (1%), and internal carotid artery (1%). Observed SAAs (nonoperatively managed group). The management of all SAAs is shown in Fig 3. There were 101 patients with 108 SAAs who were observed without operative intervention. Seventy-five patients had two CT scans that were on average 50 6 42 months apart. Twenty-six patients with 28 SAAs were lost to follow-up. SAAs were followed up for 50 6 42 months and demonstrated a slow observed mean growth rate of 0.064 6 0.18 cm/y without any ruptures (Fig 4, A). Of note,
Journal of Vascular Surgery Volume
-,
Number
Erben et al
5
-
Fig 3. Management of patients with splanchnic artery aneurysms (SAAs). NOR, Nonoperative management; OR, operative intervention.
Fig 4. A, Frequency distribution of observed growth rates. B, Frequency distribution of estimated growth rates.
10 patients had an SAA >2.0 cm in diameter. The frequency distribution of estimated growth rates is demonstrated in Fig 4, B and was calculated to be 0.086 6 0.029 cm/y. To evaluate the impact of comorbidities on growth rates, a multivariable regression analysis was conducted (Table IV). Only female sex and smoking history were found to be statistically significant. Female sex (P ¼ .02; odds ratio, 1.020; 95% confidence interval, 1.003-1.037) was associated with a faster growth rate; a history of smoking (P ¼ .04; odds ratio, 0.981; 95% confidence interval, 0.963-0.999) was associated with a slower growth rate. All other variables evaluated including age, chronic obstructive pulmonary disease, coronary artery disease, connective tissue disorder, diabetes mellitus, chronic kidney disease, history of stroke or transient ischemic attack, hypertension, congestive heart failure, peripheral arterial disease, and presence of multiple aneurysms did not contribute to the growth of SAAs.
SAAs treated (operative intervention group). Twentyfive patients with 30 SAAs underwent operative repair (Fig 2). The mean aneurysm size was 2.41 6 1.23 cm (the mean size of SAAs in the nonoperatively managed group was 1.58 6 0.56 cm; P ¼ .00001). Four of the five patients who presented with ruptured SAAs underwent repair. The 0.6-cm ruptured splenic artery pseudoaneurysm underwent coil embolization; the 3.6-cm ruptured superior mesenteric artery pseudoaneurysm underwent open bypass. For the true aneurysms that ruptured, the 2.3-cm gastroduodenal artery aneurysm underwent coil embolization, whereas the 2.3-cm splenic artery aneurysm underwent ligation with splenectomy. The patient with the 5.0-cm ruptured common hepatic artery aneurysm chose not to undergo intervention and died. All other nonruptured aneurysms underwent repair because of large size (23%), rapid growth (7%), other open abdominal surgical procedure required (7%), multiple aneurysms (3%), and desire to pursue fertility
6
Journal of Vascular Surgery
Erben et al
---
2018
Table IV. Multivariable regression analysis of the impact of comorbidities on growth rate Variable
P value (odds ratio, 95% confidence interval)
Direction of impact on SAA growth
Female sex
Increased
Smoking history
Decreased
.024 (1.02, 1.003-1.037) .041 (0.98, 0.963-0.999)
Age
d
.097 (1.00, 0.999-1.00)
COPD
d
.086 (1.02, 0.996-1.054)
CAD
d
.605 (0.99, 0.975-1.015)
Connective tissue disease
d
.903 (0.98, 0.738-1.308)
Diabetes mellitus
d
.459 (0.99, 0.975-1.012)
CKD
d
.542 (1.01, 0.988-1.023)
History of stroke or TIA
d
.730 (0.993, 0.956-1.032)
Hypertension
d
.569 (1.005, 0.988-1.023) .940 (1.00, 0.946-1.062)
CHF
d
PAD
d
.841 (0.99, 0.881-1.109)
Multiple aneurysms
d
.068 (1.01, 0.999-1.029)
CAD, Coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; PAD, peripheral arterial disease; SAA, splanchnic artery aneurysm; TIA, transient ischemic attack.
Table V. Location and mode of repair for splanchnic artery aneurysms (SAAs) Location
No.
Size, cm
Endovascular repair, No. (%)
Open repair, No. (%)
Type of open repair (No.)
12
2.44 6 1.31
9 (75)
3 (25)
Splenectomy (2), resection (1)
SMA
5
3.38 6 1.46
1 (20)
4 (80)
Bypass (4)
SMA branch
3
1.88 6 0.54
3 (75)
1 (25)
Bypass (1)
CA
3
1.43 6 0.26
2 (67)
1 (33)
Bypass (1)
CHA
2
2.95 6 0.35
2 (100)
PDA
2
2.50 6 0.30
1 (50)
GDA
1
2.3
IMA
1
0.6
Splenic
1 (100) 0 (0)
0 (0) 1 (50) 0 (0) 1 (100)
d Plication (1) d Resection (1)
CA, Celiac artery; CHA, common hepatic artery; GDA, gastroduodenal artery; IMA, inferior mesenteric artery; PDA, pancreaticoduodenal artery; SMA, superior mesenteric artery.
treatment (3%) at a mean time of 35.4 6 23.5 months from initial presentation. The repair modality is illustrated in Table V and included both endovascular coil embolization and open surgical techniques, including bypass (55%), splenectomy (18%), resection (18%), and plication (9%). Complications (Table VI) related to the procedure or operation occurred in six patients (20%), which included two in the coil embolization group: one open exploration due to hemorrhage and one splenectomy. Postoperative complications from the open surgical group included acute kidney injury in two patients that resolved without any intervention and death due to multisystem organ failure in one patient after open bypass repair of a 4.6-cm superior mesenteric artery aneurysm. Last, one superior mesenteric artery bypass was found to be occluded on follow-up, incidentally detected on crosssectional imaging, without any sequelae.
DISCUSSION Our series is the first and only one with longitudinal follow-up and growth rate calculations in patients
with SAAs. We have been able to determine the growth rate of SAAs at 0.064 6 0.18 cm/y while following their natural history. Furthermore, we have analyzed the outcomes of those SAAs that underwent operative repair, including both endovascular and open techniques. There are few data available in the literature regarding the natural history of SAAs. Most available series have a small number of patients and recommend aggressive intervention at a diameter >2.0 cm because of an anticipated high mortality rate.11,12 One aspect that remains constant in the reviews available is the lethality of SAAs in pregnant women. Studies report a maternal mortality rate of 75% and fetal mortality rate of 95% for ruptured SAA in pregnancy, thereby explaining the emphasis on early intervention.13-15 We report in our cohort that the desire to pursue fertility treatment in the presence of an SAA was an indication to undergo repair of one aneurysm, which is certainly in agreement with the reported literature.16-20
Journal of Vascular Surgery Volume
-,
Number
Erben et al
7
-
Table VI. Complications after open operative repair Outcome
No. (%)
Details
Mortality
1 (3)
Multisystem organ failure
Overall complications
6 (20)
Hemorrhage
1 (3)
Open exploration
Splenectomy
1 (3)
After embolization
Acute kidney injury
2 (7)
After embolization
Bypass occlusion
1 (3)
Time unknown
Aneurysm location. One important aspect that we encountered in our cohort was the relatively large number of celiac artery aneurysms (46% instead of <5%) and the relatively low number of splenic artery aneurysms (30% instead of >60%) that are mostly reported in the literature.12,21 We believe the reason for this selection bias is that we used cross-sectional imaging as the only imaging technique in our study to ensure that all patients were measured by the same objective imaging modality to allow fair comparison on follow-up. The majority of the series reporting splenic artery aneurysms included ultrasound as a major diagnostic tool, which could have biased our selection of patients.22-25 Furthermore, once splenic artery aneurysms are diagnosed, they undergo angiography for endovascular repair and skip the cross-sectional imaging; this negatively affects our selection of patients in using a cross-sectional imaging database search.26 Furthermore, our study includes a retrospective review of radiologic data and medical records, which could as well have influenced the selection bias because we depend on the medical record coding and dictation by the radiologist reading the cross-sectional imaging study. Observed SAAs (nonoperatively managed group). A most recent review by Corey et al2 recommended SAA surveillance and aggressive intervention in aneurysms of the pancreaticoduodenal arcade and gastroduodenal artery because these were most likely to rupture in their cohort. In our series, we were able to demonstrate a slow growth rate in SAAs of 0.064 6 0.18 cm/y and that the smallest true aneurysm rupture occurred at 2.3 cm in diameter in small-caliber arteries (splenic and gastroduodenal arteries). Unfortunately, in our series, we encountered only a handful of pancreaticoduodenal and gastroduodenal artery aneurysms (n ¼ 4), which hinders the direct correlation of these smaller caliber vessels and their higher rupture risk. Although the smallest rupture occurred at 2.3 cm, in our cohort, we observed 10 patients with SAAs >2.0 cm without any rupture. Therefore, it seems feasible to observe all SAAs with a very low risk of rupture. On multivariable regression analysis, female sex was associated with a faster growth rate, whereas a history of smoking was associated with a slower growth rate of
SAAs. Therefore, women may have a higher risk of adverse outcomes in the face of SAA, which has been reported in multiple prior studies, justifying a higher risk of mortality in the female patient with SAA.27,28 Another important factor is the known correlation between estrogen and aortic medial degeneration, which could potentially explain the risk of female patients for a worse outcome and relative protection of male patients with SAA.29,30 Interestingly, it has been shown that smoking is a risk factor for growth in the face of both an infrarenal abdominal aortic aneurysm31 and an intracranial aneurysm; however, the exact mechanism is not well understood.32,33 Our analysis demonstrated that smoking was associated with slower growth of SAAs. However, we are not sure how to interpret these results. It may be possible that these results do not apply in the growth rate of SAA. These data should be interpreted with caution. With the current data available from this study, we would recommend in patients with small SAAs yearly cross-sectional imaging follow-up because of the slow rate of growth in these SAAs. These findings, although encouraging, represent the first natural history study with growth rate calculations. Therefore, we would welcome and encourage other natural history projects to corroborate these data. SAAs treated (operative intervention group). In our review, we identified 30 SAAs in 25 patients who underwent operative repair (Fig 3). Six of these presented ruptured with two pseudoaneurysms of the splenic and superior mesenteric artery measuring 0.6 cm and 3.6 cm in diameter. There is general consensus to repair pseudoaneurysms because of their unpredictable risk of rupture at any size,34 hence their immediate repair on presentation in our cohort. All true aneurysms that underwent operative intervention had a mean size of 2.41 6 1.23 cm vs 1.58 6 0.56 cm in those patients whose SAAs were observed (P ¼ .00001). Our operative morbidity and mortality (20% and 3%, respectively) were comparable to the ones reported in the literature.2,9,11 One patient died of multisystem organ failure after a bypass for a superior mesenteric artery aneurysm of 4.6 cm in diameter. Furthermore, the operative explorations from those patients treated with endovascular techniques were expected complications. One patient underwent exploration because of hemorrhage from a ruptured gastroduodenal artery with hemodynamic instability, for whom no bleeding source was found, and a second patient with a necrotic spleen required splenectomy. The complications in those patients who underwent open repair included acute kidney injury, which resolved without any intervention, and an occluded bypass at an unknown time after operation without any negative sequelae.
8
Journal of Vascular Surgery
Erben et al
---
We believe that a strength of our natural history report using the radiologic database for SAAs is that it offers a true representation of all diagnosed patients with this disease entity. Most are asymptomatic, small aneurysms with low risk of rupture. Previous reports are derived from surgical databases, and therefore most of these patients are already preselected for operative intervention.2,6,7 By observing all patients with abnormalities on CT scans, we are able to assess the natural history of SAAs, including incidentally found SAAs. Study limitations. We have encountered several limitations with the current project, which include the relatively small number of aneurysms analyzed and the retrospective nature of our review. Furthermore, to objectively assess growth rates, we used cross-sectional imaging in the form of CT to assess the size of the vessels in question. Therefore, a large proportion of patients diagnosed by other imaging modalities have not been taken into consideration, which could have clearly strengthened the study by the sheer number of patients. However, because of unpredictability of ultrasound measurements, these were not considered. Next, if codes for SAA diagnosis were not entered properly within the electronic medical records of patients, these potential patients with SAA were not picked up at the time of data collection.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
CONCLUSIONS Our report regarding the natural history of SAAs is able to calculate growth rates for SAAs. The growth rate of SAAs is very slow (0.064 6 0.18 cm/y), and therefore it seems reasonable to observe asymptomatic patients with a small to medium-sized SAA. In our series, the smallest ruptured SAA occurred at 2.3 cm in diameter. If operative intervention should be needed, both open and endovascular options can be considered.
13.
14.
15.
AUTHOR CONTRIBUTIONS Conception and design: YE, AB, JE Analysis and interpretation: YE, AB, SR, YL, JR, HM, BZ, JE Data collection: YE, AB, SR, HM Writing the article: YE, AB, BZ, JE Critical revision of the article: YE, AB, SR, YL, JR, HM, BZ, JE Final approval of the article: YE, AB, SR, YL, JR, HM, BZ, JE Statistical analysis: AB, YL, JR Obtained funding: Not applicable Overall responsibility: JE YE and AB contributed equally to this article and share first authorship.
16.
17.
18.
19.
20.
REFERENCES 1. Bedford PD, Lodge B. Aneurysm of the splenic artery. Gut 1960;1:312-20. 2. Corey MR, Ergul EA, Cambria RP, English SJ, Patel VI, Lancaster RT, et al. The natural history of splanchnic artery
21. 22.
2018
aneurysms and outcomes after operative intervention. J Vasc Surg 2016;63:949-57. Abbas MA, Fowl RJ, Stone WM, Panneton JM, Oldenburg WA, Bower TC, et al. Hepatic artery aneurysm: factors that predict complications. J Vasc Surg 2003;38: 41-5. Abbas MA, Stone WM, Fowl RJ, Gloviczki P, Oldenburg WA, Pairolero PC, et al. Splenic artery aneurysms: two decades experience at Mayo clinic. Ann Vasc Surg 2002;16:442-9. Saltzberg SS, Maldonado TS, Lamparello PJ, Cayne NS, Nalbandian MM, Rosen RJ, et al. Is endovascular therapy the preferred treatment for all visceral artery aneurysms? Ann Vasc Surg 2005;19:507-15. Stone WM, Abbas M, Cherry KJ, Fowl RJ, Gloviczki P. Superior mesenteric artery aneurysms: is presence an indication for intervention? J Vasc Surg 2002;36:234-7; discussion: 237. Stone WM, Abbas MA, Gloviczki P, Fowl RJ, Cherry KJ. Celiac arterial aneurysms: a critical reappraisal of a rare entity. Arch Surg 2002;137:670-4. Tetreau R, Beji H, Henry L, Valette PJ, Pilleul F. Arterial splanchnic aneurysms: presentation, treatment and outcome in 112 patients. Diagn Interv Imaging 2016;97: 81-90. Shukla AJ, Eid R, Fish L, Avgerinos E, Marone L, Makaroun M, et al. Contemporary outcomes of intact and ruptured visceral artery aneurysms. J Vasc Surg 2015;61:1442-7. Rizzo JA, Coady MA, Elefteriades JA. Procedures for estimating growth rates in thoracic aortic aneurysms. J Clin Epidemiol 1998;51:747-54. Panayiotopoulos YP, Assadourian R, Taylor PR. Aneurysms of the visceral and renal arteries. Ann R Coll Surg Engl 1996;78: 412-9. van Rijn MJ, Ten Raa S, Hendriks JM, Verhagen HJ. Visceral aneurysms: old paradigms, new insights? Best Pract Res Clin Gastroenterol 2017;31:97-104. Hunsaker DM, Turner S, Hunsaker JC 3rd. Sudden and unexpected death resulting from splenic artery aneurysm rupture: two case reports of pregnancy-related fatal rupture of splenic artery aneurysm. Am J Forensic Med Pathol 2002;23:338-41. Veluppillai C, Perreve S, de Kerviler B, Ducarme G. [Splenic arterial aneurysm and pregnancy: a review]. Presse Med 2015;44:991-4. Stanley JC, Wakefield TW, Graham LM, Whitehouse WM Jr, Zelenock GB, Lindenauer SM. Clinical importance and management of splanchnic artery aneurysms. J Vasc Surg 1986;3:836-40. Benali M, Charrada H, Bouassida M, Bahloul A, Jmal K, Dhouib F, et al. [Splenic artery aneurysm rupture in late pregnancy: a case report]. Ann Fr Anesth Reanim 2013;32: 721-2. Lang W, Strobel D, Beinder E, Raab M. Surgery of a splenic artery aneurysm during pregnancy. Eur J Obstet Gynecol Reprod Biol 2002;102:215-6. Shahabi S, Jani J, Masters L, Cobin L, Greindl J. Spontaneous rupture of a splenic artery aneurysm in pregnancy: report of two cases. Acta Chir Belg 2000;100:231-3. Arrieta FM, Muguerza J, Garcia Sancho L, Ayuso M, Rustarazu M, Valenzuela P. Rupture of splenic artery aneurysm during pregnancy and posterior evolution of gestation. Zentralbl Gynakol 2000;122:579-80. Cressey D, Reid MF. Splenic artery aneurysm rupture in pregnancy. Int J Obstet Anesth 1996;5:103-4. Pasha SF, Gloviczki P, Stanson AW, Kamath PS. Splanchnic artery aneurysms. Mayo Clin Proc 2007;82:472-9. Fransen H, Kubale R, Wurche KD, Kalahne A. [Noninvasive diagnosis of splenic artery aneurysm]. Rofo 1989;151:532-5.
Journal of Vascular Surgery Volume
-,
Number
Erben et al
9
-
23. Davis T, Minardi J, Knight J, Larrabee H, Schaefer G. Ruptured splenic artery aneurysm: rare cause of shock diagnosed with bedside ultrasound. West J Emerg Med 2015;16:762-5. 24. Lo WL, Mok KL. Ruptured splenic artery aneurysm detected by emergency ultrasoundda case report. Crit Ultrasound J 2015;7:26. 25. Dave SP, Reis ED, Hossain A, Taub PJ, Kerstein MD, Hollier LH. Splenic artery aneurysm in the 1990s. Ann Vasc Surg 2000;14:223-9. 26. Hogendoorn W, Lavida A, Hunink MG, Moll FL, Geroulakos G, Muhs BE, et al. Open repair, endovascular repair, and conservative management of true splenic artery aneurysms. J Vasc Surg 2014;60:1667-76.e1. 27. Ha JF, Phillips M, Faulkner K. Splenic artery aneurysm rupture in pregnancy. Eur J Obstet Gynecol Reprod Biol 2009;146:133-7. 28. Lynch MJ, Woodford NW. Rupture of a splenic artery aneurysm during pregnancy with maternal and foetal death: a case report. Med Sci Law 2008;48:342-5. 29. Crawford JD, Hsieh CM, Schenning RC, Slater MS, Landry GJ, Moneta GL, et al. Genetics, pregnancy, and aortic degeneration. Ann Vasc Surg 2016;30:158.e5-9.
30. Tripathi R, Sainathan S, Ziganshin BA, Elefteriades JA. Thoracic aortic aneurysm from chronic antiestrogen therapy. Int J Angiol 2017;26:60-3. 31. Enriquez-Vega ME, Solorio-Rosete HF, Cossio-Zazueta A, Bizueto-Rosas H, Cruz-Castillo JE, Iturburu-Enriquez A. [Early detection of abdominal aortic aneurysm in risk population]. Rev Med Inst Mex Seguro Soc 2015;53(Suppl 1):S100-3. 32. Backes D, Rinkel GJ, Greving JP, Velthuis BK, Murayama Y, Takao H, et al. ELAPSS score for prediction of risk of growth of unruptured intracranial aneurysms. Neurology 2017;88: 1600-6. 33. Kim HW, Stansfield BK. Genetic and epigenetic regulation of aortic aneurysms. Biomed Res Int 2017;2017:7268521. 34. Zyromski NJ, Vieira C, Stecker M, Nakeeb A, Pitt HA, Lillemoe KD, et al. Improved outcomes in postoperative and pancreatitis-related visceral pseudoaneurysms. J Gastrointest Surg 2007;11:50-5.
Submitted Oct 17, 2017; accepted Dec 30, 2017.