- Email: [email protected]
Causes and outcomes of finger ischemia in hospitalized patients in the intensive care unit
From the Western Vascular Society
Causes and outcomes of finger ischemia in hospitalized patients in the intensive care unit Gregory J. Landry, MD, Courtney J. Mostul, Daniel S. Ahn, BS, Bryant J. McLafferty, Timothy K. Liem, MD, Erica L. Mitchell, MD, Enjae Jung, MD, Cherrie Z. Abraham, MD, Amir F. Azarbal, MD, Robert B. McLafferty, MD, and Gregory L. Moneta, MD, Portland, Ore
ABSTRACT Objective: Vascular surgeons may be consulted to evaluate hospitalized patients with finger ischemia. We sought to characterize causes and outcomes of finger ischemia in intensive care unit (ICU) patients. Methods: All ICU patients who underwent evaluation for finger ischemia from 2008 to 2015 were reviewed. All were evaluated with finger photoplethysmography. The patients’ demographics, comorbidities, ICU care (ventilator status, arterial lines, use of vasoactive medications), finger amputations, and survival were also recorded. ICU patients were compared with concurrently evaluated non-ICU inpatients with finger ischemia. Results: There were 98 ICU patients (55 male, 43 female) identified. The mean age was 57.1 6 16.8 years. Of these patients, 42 (43%) were in the surgical ICU and 56 (57%) in the medical ICU. Seventy (72%) had abnormal findings on finger photoplethysmography, 40 (69%) unilateral and 30 (31%) bilateral. Thirty-six (37%) had ischemia associated with an arterial line. Twelve (13%) had concomitant toe ischemia. Eighty (82%) were receiving vasoactive medications at the time of diagnosis, with the most frequent being phenylephrine (55%), norepinephrine (47%), ephedrine (31%), epinephrine (26%), and vasopressin (24%). Treatment was with anticoagulation in 88 (90%; therapeutic, 48%; prophylactic, 42%) and antiplatelet agents in 59 (60%; aspirin, 51%; clopidogrel, 15%). Other frequently associated conditions included mechanical ventilation at time of diagnosis (37%), diabetes (34%), peripheral arterial disease (32%), dialysis dependence (31%), cancer (24%), and sepsis (20%). Only five patients (5%) ultimately required finger amputation. The 30-day, 1-year, and 3-year survival was 84%, 69%, and 59%. By Cox proportional hazards modeling, cancer (hazard ratio, 2.4; 95% confidence interval, 1.1-5.6; P ¼ .035) was an independent predictor of mortality. There were 50 concurrent non-ICU patients with finger ischemia. Non-ICU patients were more likely to have connective tissue disorders (26% vs 13%; P ¼ .05) and hyperlipidemia (42% vs 24%; P ¼ .03) and to undergo finger amputations (16% vs 5%; P ¼ .03). Conclusions: Finger ischemia in the ICU is frequently associated with the presence of arterial lines and the use of vasopressor medications, of which phenylephrine and norepinephrine are most frequent. Anticoagulation or antiplatelet therapy is appropriate treatment. Whereas progression to amputation is rare, patients with finger ischemia in the ICU have a high rate of mortality, particularly in the presence of cancer. Non-ICU patients hospitalized with finger ischemia more frequently require finger amputations, probably because of more frequent connective tissue disorders. (J Vasc Surg 2018;68:1499-504.) Keywords: Finger; Vasospasm; Ischemia; Intensive care unit
Vascular surgeons may be consulted to evaluate patients in the intensive care unit (ICU) with digital ischemia. Several factors can lead to digital ischemia in these patients, including the severity of underlying illness From the Division of Vascular Surgery, Knight Cardiovascular Institute, Oregon Health & Science University. Author conflict of interest: none. Presented at the Thirty-second Annual Meeting of the Western Vascular Society, Blaine, Wash, September 23-26, 2017. Correspondence: Gregory J. Landry, MD, Division of Vascular Surgery, Knight Cardiovascular Institute, Oregon Health & Science University, 3181 SW Sam Jackson Park Rd, Mail Code OP11, Portland, OR 97239 (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.2018.01.050
(eg, sepsis,1,2 heart failure, malignant disease,3,4 renal failure,5 and diabetes6), use of vasoactive medications causing peripheral vasoconstriction,7-11 and use of radial arterial lines for continuous hemodynamic monitoring.12,13 Finger ischemia can be the source of great anxiety for critically ill patients, families, and care providers. Whereas these patients have multiple critical health issues by virtue of requiring ICU care, the hands are very visible, and finger abnormalities are a prominent outward sign of the patient’s illness. Although surgical intervention is not frequently required, vascular surgeons are often consulted to provide guidance as a vascular “specialist.” The immediate and long-term outcomes of patients in the ICU with finger ischemia are not well characterized and form the basis of this report. We sought to determine factors associated with digital ischemia in ICU patients and 1499
1500
Journal of Vascular Surgery
Landry et al
November 2018
to determine short- and long-term outcomes of the affected digits as well as short- and long-term survival of patients who develop finger ischemia in the ICU.
ARTICLE HIGHLIGHTS d
d
METHODS All patients in either medical or surgical ICUs at a single tertiary care facility (Oregon Health & Science University, Portland, Oregon) developing finger ischemia and who had vascular surgery referral from 2008 to 2015 were identified and evaluated. Patients were identified by querying the billing code for finger photoplethysmography (PPG) performed by an Intersocietal Accreditation Commission-approved vascular laboratory at this facility. The hospital records of all patients were evaluated, with data abstracted from the patients’ electronic medical record (Epic; Epic Systems Corporation, Madison, Wisc). Historical demographic data were recorded, including age, sex, body mass index, history of diabetes, hypertension, hyperlipidemia, peripheral arterial disease, chronic obstructive pulmonary disease, coronary artery disease, cancer, renal insufficiency or failure, and connective tissue disorders. We evaluated reason for ICU admission as well as additional concomitant ICU factors that could contribute to finger ischemia, such as the presence of radial arterial lines, mechanical ventilation, need for hemodialysis, and use of vasoactive medications. Type and duration of vasoactive medications were also recorded. Additional medications evaluated included anticoagulants, antiplatelet agents, antibiotics, beta blockers, and statin medications. Symptoms prompting referral and results of the finger PPG evaluation were also recorded. Active warming was used as needed to bring finger temperatures to at least room temperature. Abnormal findings included dampening or complete flattening of the PPG waveforms. Outcomes evaluated included the need for partial or complete finger amputation and overall survival. A concurrent control group of hospitalized, non-ICU patients undergoing finger PPG evaluation was also evaluated with the same information recorded. All categorical variables are presented as numbers and percentages. Continuous variables are presented as mean 6 standard deviation for normally distributed data and median with range for non-normally distributed data. Categorical data are compared with Fisher exact test or c2 test as appropriate. Continuous variables are evaluated with Student t-test. Kaplan-Meier analysis of survival was performed, and multivariate analysis of survival was performed using Cox proportional hazards modeling. The study was approved by the Institutional Review Board at Oregon Health & Science University. Because of the retrospective nature of the study and use of deidentified data, informed consent was not required for the enrolled patients.
d
Type of Research: Retrospective cohort study Take Home Message: In 98 intensive care unit patients with finger ischemia, the most frequent causes were the use of vasoactive medications and arterial lines. Finger amputations were needed in 5%, and finger ischemia was a marker for decreased survival, particularly so in patients with cancer. Recommendation: This study suggests that intensive care unit patients with finger ischemia have a low rate of finger amputation, but they have increased mortality, particularly if they have cancer.
RESULTS During the study period, 552 patients underwent finger PPG evaluations for upper extremity or digital ischemia. Of these, 404 were outpatient evaluations or duplicate studies and were excluded. Of the remaining patients, 98 were ICU patients and 50 were non-ICU inpatients. Of the 98 ICU patients, 55 were male and 43 were female. The mean age was 57.1 6 16.8 years. Of these patients, 42 were in the surgical ICU and 56 in the medical ICU. The symptoms were unilateral in 68 patients and bilateral in 30. Presenting symptoms included discoloration in all patients, with sensory findings (pain or paresthesias) in 47 (48%) and the presence of digital ulcerations or gangrene in 3 (3%). During the same period, 50 non-ICU patients with finger ischemia were identified to serve as a control group. Demographics of the ICU and non-ICU patients are included in Table I. In comparison to the ICU patients, non-ICU patients more frequently had hyperlipidemia (24 [25%] vs 21 [42%]; P ¼ .03) and connective tissue disorders (13 [13%] vs 13 [26%]; P ¼ .05), with a trend for increased diabetes (33 [34%] vs 23 [46%]; P ¼ .11) and dialysis (30 [31%] vs 21 [42%]; P ¼ .17). No other significant demographic differences were noted. Among ICU patients, findings on finger PPG studies were normal in 28 patients (29%). Unilateral abnormal findings were present in 40 (41%) patients, with the median number of fingers affected 3.5, mode 5. Bilateral abnormalities were noted in 30 patients (31%), with a median of 7.5 fingers affected, mode 10. Thus, when a hand is affected, it most frequently affected all digits in that hand. In comparison, non-ICU patients had a trend for more normal findings on PPG studies (n ¼ 21 [42%]) and unilateral PPG abnormalities (n ¼ 21 [42%]) and fewer bilateral abnormalities (n ¼ 8 [16%]; P ¼ .10). For each patient, the average digital pressure for all abnormal fingers was recorded. Among ICU patients, the digital pressures were lower than for non-ICU patients (54.7 6 45.9 vs 74.1 6 40.6; P ¼ .02). Correspondingly, digital brachial indices in ICU patients were also
Journal of Vascular Surgery
Landry et al
1501
Volume 68, Number 5
Table I. Demographic factors for intensive care unit (ICU) and non-ICU patients with finger ischemia Factor
ICU (n ¼ 98)
Non-ICU (n ¼ 50)
P value
Table II. Vasoactive medication frequency and duration in intensive care unit (ICU) patients with finger ischemia Patients (n ¼ 98), No. (%)
Medication
No. of days, mean 6 SD
Sex, M/F
55 (56)/43 (44) 24 (48)/26 (52)
.35
Norepinephrine
46 (47)
4.3 6 4.8
Age, years
57.1 6 16.8
56.6 6 17.4
.86
Phenylephrine
54 (55)
2.5 6 3.3
Smoking history
58 (59)
28 (56)
.78
Ephedrine
30 (31)
1.5 6 0.8
Hypertension
65 (66)
33 (66)
.97
Epinephrine
25 (26)
2.9 6 2.6
Diabetes
33 (34)
23 (46)
.11
Vasopressin
24 (24)
2.5 6 1.6
Peripheral vascular disease
31 (32)
19 (38)
.439
Dopamine
7 (7)
2.4 6 1.5
Dobutamine
6 (6)
2.3 6 1.0
Coronary artery disease
30 (31)
16 (33)
.80
SD, Standard deviation.
Congestive heart failure
19 (20)
8 (16)
.60
Hyperlipidemia
24 (25)
21 (42)
.03
Table III. Medical therapy for patients with finger ischemia
Cancer
23 (24)
10 (20)
.63
Medications
Dialysis
ICU (n ¼ 98), No. (%)
Non-ICU (n ¼ 50), No. (%)
P value
30 (31)
21 (42)
.17
Antibiotics
83 (85)
31 (62)
.002
Connective tissue disease
13 (13)
13 (26)
.05
Anticoagulation
88 (90)
33 (66)
<.001
Therapeutic
47 (48)
17 (34)
.19
Chronic obstructive pulmonary disease
12 (12)
6 (12)
.97
Prophylactic
41 (42)
16 (32)
.25 .80
Body mass index, kg/m2
27.4 6 9.9
26.7 6 10.0
.71
Categorical variables are presented as number (%). Continuous variables are presented as mean 6 standard deviation.
Antiplatelet
59 (60)
29 (58)
Aspirin
50 (51)
28 (56)
.57
15 (15)
7 (14)
.83
Beta blocker
71 (72)
29 (58)
.08
Statin
46 (47)
19 (38)
.30
Clopidogrel
ICU, Intensive care unit.
lower (0.45 6 0.36 vs 0.59 6 0.33; P ¼ .030). The lower values noted in ICU patients are likely to represent a higher number of patients with acute illnesses, resulting in completely flattened waveforms with undetectable digital pressures, most of which resolved with systemic improvement, in comparison to the non-ICU patients. Radial arterial lines were present in 36 patients (37%). Of those with arterial lines, the majority had unilateral symptoms (n ¼ 24 [66%]); however, 12 (33%) had bilateral symptoms. This was not statistically different from the distribution noted in patients without arterial lines. PPG results in patients with arterial lines were normal in 10 (28%), unilaterally abnormal in 14 (39%), and bilaterally abnormal in 12 (33%). Twenty-four patients with arterial lines underwent upper extremity arterial duplex ultrasound examination, with documented radial artery thrombosis in 16 (67%) but no thrombosis noted in 8 (33%). No finger amputations were required among patients with radial arterial lines. A number of concomitant ICU conditions were identified, including mechanical ventilation in 36 patients (37%), hemodialysis in 30 (31%), and septic shock in 19 (20%). Twelve patients (13%) had concomitant toe ischemia. However, the most frequent ICU-related factor was the use of intravenous vasoactive medications, present in 80 (82%) patients. The most frequent vasoactive medications and the duration of use are noted in Table II. Phenylephrine (n ¼ 54 [55%]) and
norepinephrine (n ¼ 46 [47%]) were the most frequently used medications. Norepinephrine was used for the longest duration (4.2 days), with the remaining medications used for 3 days or less. Most patients were taking more than one vasoactive agent, with the total number of vasoactive agents as follows: none in 18 patients (18%), one in 21 patients (21%), two in 23 patients (24%), three in 22 patients (22%), four in 11 patients (11%), and five in 3 patients (3%). Additional medications are listed in Table III. Antibiotic use was much more common in ICU patients than in non-ICU patients (85% vs 62%; P ¼ .02). Therapeutic and prophylactic anticoagulation, antiplatelet medications (aspirin or clopidogrel), beta blockers, and statins were commonly used in both ICU and non-ICU patients. Overall anticoagulation rates were higher in ICU patients (90% vs 66%; P < .001); however, there were no significant differences noted when anticoagulants were subdivided into therapeutic and prophylactic doses. Among patients treated with therapeutic anticoagulation, most were already being treated therapeutically for other thrombotic events or risk factors (74% ICU vs 71% non-ICU; P ¼ NS), with the remainder started on anticoagulation specifically for their hand ischemia. Among the ICU patients, only five went on to digital amputations. In comparison, amputations were performed in eight non-ICU patients (5% vs 16%; P ¼ .03).
1502
Journal of Vascular Surgery
Landry et al
November 2018
Fig 1. Survival of patients with finger ischemia by intensive care unit (ICU) status. S.E., Standard error.
Non-ICU patients more frequently had a history of prior finger amputations than ICU patients (6 [12%] vs 3 [3%]; P ¼ .06). ICU patients undergoing amputation underwent amputation of 1.6 6 0.6 fingers compared with 1.1 6 0.4 (P ¼ .08) in non-ICU patients. The ICU length of stay in ICU patients was 11.8 6 10.8 days. Overall hospital length of stay was 20.5 6 14.6 days, significantly greater than overall hospital length of stay in non-ICU patients (7.3 6 4.5 days; P < .001). Survival of ICU vs non-ICU patients is shown in Fig 1. The 30-day survival was significantly lower for the ICU patients (84% vs 96%; P ¼ .04). Survival at 1 year, 3 years, and 5 years after a finger ischemic event was 69%, 59%, and 51% in ICU patients and 79%, 63%, and 42% in non-ICU patients (P ¼ .798). Among ICU patients, multivariate analysis with Cox proportional hazards modeling revealed cancer as the only predictor of mortality (hazard ratio, 2.4; 95% confidence interval, 1.1-5.6; P ¼ .035). Among non-ICU patients, the only significant multivariate predictor of mortality was increasing age (hazard ratio, 1.06; 95% confidence interval, 1.00-1.11; P ¼ .023), with the mean age of 51.4 6 18.1 years for survivors and 63.2 6 14.4 years for nonsurvivors. Among patients with finger amputations, including both ICU and non-ICU patients, survival at 1 year, 3 years, and 5 years was 83%, 47%, and 17% compared with 75%, 66%, and 61% of those not undergoing amputation at the same time points (P ¼ .168; Fig 2). There were too few amputation events to stratify for ICU and non-ICU patients, although among the ICU patients, the longest survival of a patient undergoing a finger amputation was 21 months.
Fig 2. Survival of patients with finger ischemia by finger amputation status. S.E., Standard error.
DISCUSSION Patients in the ICU frequently have hemodynamic instability requiring the use of vasoactive medications. However, a potential adverse sequela of hemodynamic support is reduced end-organ perfusion. The extremities are particularly vulnerable, and several reports of upper and lower extremity digital malperfusion in patients receiving vasopressor agents have been published.7-11 Whereas bilateral involvement is most frequent, asymmetric unilateral involvement has been described as was seen in our series. Previous studies have shown pre-existing vascular disease, concomitant use of multiple vasopressor agents, and prolonged periods of hypotension to be risk factors for digital ischemia.10 Most reports in the medical literature have identified the association between norepinephrine and risk of digital ischemia.7-11 Norepinephrine was used in almost half the patients in our series and was the vasopressor agent used for the longest duration. Of the five ICU patients who required digital amputation, two were taking norepinephrine and one each was taking phenylephrine, vasopressin, and epinephrine. In a systematic review from the Scandinavian Society of Anaesthesiology and Intensive Care Medicine, norepinephrine was identified as the first-line vasopressor of choice, with moderate quality of evidence in patients with shock in general and septic shock and low-quality evidence for cardiogenic or hypovolemic shock.14 Whereas a maximal tolerable dose of norepinephrine has not been definitively established, it is known that a high dose and duration of norepinephrine are associated with a high peripheral vasoconstriction risk, and an optimal dose of 0.2 to 1.3 mg/kg/min is
Journal of Vascular Surgery
Landry et al
1503
Volume 68, Number 5
associated with a reduced risk of end-organ ischemia.7 Accidental arterial injections causing hand ischemia have been reported with other vasoactive agents, such as ephedrine15 and epinephrine.16 No intra-arterial injections occurred in the current series. Digital necrosis due to high-dose intravenous infusion of epinephrine and dopamine has been reported.17 The management of digital ischemia in ICU patients is not well established. Overall, prognosis is clearly related to the severity of underlying disease, and minimizing the dosage and duration of vasopressor use as much as possible is prudent. The use of anticoagulation or antiplatelet therapy is not well established, although these patients are often already taking these medications for other indications. In the current series, 60% of patients were receiving antiplatelet therapy and 90% either therapeutic or prophylactic anticoagulation at the time of diagnosis. In cases of presumed arterial thrombosis, these treatments are warranted; however, in cases of severe vasospasm, the addition of these agents is of questionable benefit. Finger and hand ischemia secondary to radial arterial lines has been previously described in multiple reports. Valentine et al13 described eight patients during a 5-year period. Five patients underwent surgical intervention for radial artery thrombosis, two were treated with anticoagulation, and one was treated conservatively. Rethrombosis after surgical repair with patch angioplasty was common, and all patients treated surgically required digital amputation, whereas no amputations were required in patients treated nonsurgically. Garg et al12 identified 30 patients with complications related to radial arterial lines, including 15 thrombotic episodes, 10 pseudoaneurysms, and 5 abscesses. All thrombotic cases underwent surgical thrombectomy, with three developing rethrombosis. No patients required digital amputations; however, in-hospital mortality of the cohort was 37%, reflecting the severity of illness of these patients. No amputations were required in our series; however, 30-day mortality was only 16%, probably reflecting the fact that our series included all patients with symptoms of digital ischemia, even if only mild discoloration, which may portend a more favorable prognosis than for those who require surgical intervention because of ischemia severity. Operations due to radial artery occlusion were rarely required. One patient underwent radial thrombectomy. One patient initially admitted for bilateral lower extremity ischemia due to thromboembolism subsequently developed right upper extremity ischemia due to axillary and brachial artery emboli requiring thromboembolectomy. One patient undergoing chemotherapy for lymphoma was found to have axillary and brachial artery occlusion in a pattern typical for temporal arteritis along with ulnar artery occlusion and underwent axillary to radial artery bypass. One patient admitted with
bacterial endocarditis developed left upper extremity ischemia while in the ICU due to a brachial artery embolus, for which she underwent embolectomy. The rarity of finger amputations in the ICU group probably reflects primarily a vasospastic etiology. Despite the occasional severity of presentation, the intense vasospasm brought on by the patient’s underlying critical illness in conjunction with vasoactive agents is typically reversible as the patient’s overall condition improves and vasoactive agents are withdrawn. With few exceptions, digital ischemia secondary to vasospasm portends a good prognosis with respect to tissue loss.18 In contrast, the higher rate of finger amputation in non-ICU patients is likely to reflect an association with fixed digital artery occlusion due to disorders such as chronic hemodialysis, diabetes, and connective tissue disorders, which are more strongly associated with subsequent tissue loss.5,18,19 The poor survival of patients with digital ischemia is noteworthy. Although it would be expected that ICU patients have decreased 30-day survival compared with non-ICU patients, mortality rates beyond 30 days were not different between ICU and non-ICU patients. Whereas the ICU patients had greater acuity of illness, non-ICU patients had a number of more frequent comorbidities, such as dialysis dependence, connective tissue disorders, hyperlipidemia, and diabetes, which can be associated with decreased long-term survival. This indicates that regardless of etiology, finger ischemia is a poor prognostic indicator for survival. The need for finger amputation was not associated with decreased survival, but this is likely due to the small number of patients in the amputation group and probably reflects a type II error. This study has several limitations. The patients were identified by billing codes for PPG. This study is routinely performed in all patients evaluated for finger ischemia in the ICU, so we believe this captures all of the relevant patients during this period. Nonetheless, it is possible that some patients with finger ischemia were not evaluated by the vascular surgery service and did not undergo PPG evaluation. Presumably, though, these would be patients with milder symptoms, as patients with more severe symptoms routinely receive vascular surgery consultation. Alternatively, these could be patients whose systemic illness was so severe and life expectancy so low that finger ischemia was not addressed. Our database includes only patients with manifestations of digital ischemia. Most ICU patients taking vasoactive agents do not develop digital ischemia and therefore are not included in our database. We therefore cannot determine the incidence of digital ischemia in patients taking vasoactive agents or comment on the relative safety of one vasoactive agent vs another with respect to this complication. Furthermore, the retrospective nature of the study made it difficult to determine the dosages of the vasoactive
1504
Journal of Vascular Surgery
Landry et al
November 2018
agents at the time of symptom manifestation because dosages were continuously adjusted in response to patients’ needs. We therefore are unable to determine whether there was a threshold dose that was more likely to lead to ischemic symptoms; however, this is a question that we plan to evaluate prospectively. The data regarding arterial lines are based on the presence of an arterial line at the time of consultation. It is possible that patients had a prior arterial line that was unknown to us and that the number of patients with arterial line-associated ischemia is underestimated.
6.
7.
8.
9.
CONCLUSIONS Although it is a source of anxiety for patients, families, and care providers, digital ischemia in ICU patients rarely leads to finger amputation, with finger recovery associated with patients’ recovery from their critical illness. Conservative therapy is indicated in most cases. Digital ischemia, however, is a poor prognostic sign, associated with high rates of both early and late mortality.
AUTHOR CONTRIBUTIONS Conception and design: GL, TL, EM, EJ, CA, AA, RM, GM Analysis and interpretation: GL Data collection: GL, CM, DA, BM Writing the article: GL, CM, DA, BM Critical revision of the article: GL, TL, EM, EJ, CA, AA, RM, GM Final approval of the article: GL, CM, DA, BM, TL, EM, EJ, CA, AA, RM, GM Statistical analysis: GL Obtained funding: Not applicable Overall responsibility: GL
REFERENCES 1. Johansen K, Hansen ST. Symmetrical peripheral gangrene (purpura fulminans) complicating pneumococcal sepsis. Am J Surg 1991;165:642-5. 2. Davis MD, Dy KM, Nelson S. Presentation and outcome of purpura fulminans associated with peripheral gangrene in 12 patients at Mayo Clinic. J Am Acad Dermatol 2007;57:944-56. 3. Chow SF, McKenna CH. Ovarian cancer and gangrene of the digits: case report and review of the literature. Mayo Clin Proc 1996;71:253-8. 4. Kopterides P, Tsavaris N, Tzioufas A, Pikazis D, Lazaris A. Digital gangrene and Raynaud’s phenomenon as complications of lung adenocarcinoma. Lancet Oncol 2004;5:549. 5. Yeager RA, Moneta GL, Edwards JM, Landry GJ, Taylor LM, McConnell DB, et al. Relationship of hemodialysis access to
10.
11.
12.
13. 14.
15.
16.
17.
18.
19.
finger gangrene in patients with end-stage renal disease. J Vasc Surg 2002;36:245-9. Ahmadi N, Hajsadeghi F, Gul K, Leibfried M, DeMoss D, Lee R, et al. Vascular function measured by fingertip thermal reactivity is impaired in patients with metabolic syndrome and diabetes mellitus. J Clin Hypertens 2009;11:678-84. Shin JY, Roh SG, Lee NH, Yang KM. Ischemic necrosis of upper lip, and all fingers and toes after norepinephrine use. J Craniofac Surg 2016;27:453-4. Simman R, Phavixay L. Bilateral toe necrosis resulting from norepinephrine bitartrate usage. Adv Skin Wound Care 2013;26:254-6. Chuang SS. Finger ischemia secondary to the synergistic agonist effect of norepinephrine and ergonovine and in a burn patient. Burns 2003;29:92-4. Daroca-Perez R, Carrascosa MF. Digital necrosis; a potential risk of high-dose norepinephrine. Ther Adv Drug Saf 2017;8: 259-61. McAllister SF, Cleary JD. Blue digits: cyanosis resulting from norepinephrine treatment. Ann Pharmacother 2005;39: 383-4. Garg K, Howell BW, Saltzberg SS, Berland TL, Mussa FF, Maldonado TS, et al. Open surgical management of complications from indwelling radial artery catheters. J Vasc Surg 2013;58:1325-30. Valentine RJ, Modrall JG, Clagett GP. Hand ischemia after radial artery cannulation. J Am Coll Surg 2005;201:18-22. Møller MH, Claudius C, Junttila E, Haney M, OscarssonTibblin A, Haavind A, et al. Scandinavian SSAI clinical practice guideline on choice of first-line vasopressor for patients with acute circulatory failure. Acta Anaesthesiol Scand 2016;60:1347-66. Kozar S, Kurnik G. Accidental intra arterial injection of ephedrine: what about the treatment with nitroglycerin? J Anesth Clin Res 2015;6:2. Velissariou I, Cottrell S, Berry K, Wilson B. Management of adrenaline (epinephrine) induced digital ischaemia in children after accidental injection from an EpiPen. Emerg Med J 2004;21:387-8. Reyes AJ, Ramcharan K, Harnarayan P, Mooteeram J. Symmetrical digital gangrene after a high dose intravenous infusion of epinephrine and dopamine following resuscitation from cardiac arrest. BMJ Case Rep 2016;2016. bcr2016217977. Landry GJ, Edwards JM, McLafferty RB, Taylor LM, Porter JM. Long-term outcome of Raynaud’s syndrome in a prospectively analyzed patient cohort. J Vasc Surg 1996;23:76-86. Landry GJ, McClary A, Liem TK, Mitchell EL, Azarbal AF, Moneta GL. Factors affecting healing and survival after finger amputations in patients with digital artery occlusive disease. Am J Surg 2013;205:566-70.
Submitted Nov 4, 2017; accepted Jan 22, 2018.
Causes and outcomes of finger ischemia in hospitalized patients in the intensive care unit Gregory J. Landry, MD, Courtney J. Mostul, Daniel S. Ahn, BS, Bryant J. McLafferty, Timothy K. Liem, MD, Erica L. Mitchell, MD, Enjae Jung, MD, Cherrie Z. Abraham, MD, Amir F. Azarbal, MD, Robert B. McLafferty, MD, and Gregory L. Moneta, MD, Portland, Ore
ABSTRACT Objective: Vascular surgeons may be consulted to evaluate hospitalized patients with finger ischemia. We sought to characterize causes and outcomes of finger ischemia in intensive care unit (ICU) patients. Methods: All ICU patients who underwent evaluation for finger ischemia from 2008 to 2015 were reviewed. All were evaluated with finger photoplethysmography. The patients’ demographics, comorbidities, ICU care (ventilator status, arterial lines, use of vasoactive medications), finger amputations, and survival were also recorded. ICU patients were compared with concurrently evaluated non-ICU inpatients with finger ischemia. Results: There were 98 ICU patients (55 male, 43 female) identified. The mean age was 57.1 6 16.8 years. Of these patients, 42 (43%) were in the surgical ICU and 56 (57%) in the medical ICU. Seventy (72%) had abnormal findings on finger photoplethysmography, 40 (69%) unilateral and 30 (31%) bilateral. Thirty-six (37%) had ischemia associated with an arterial line. Twelve (13%) had concomitant toe ischemia. Eighty (82%) were receiving vasoactive medications at the time of diagnosis, with the most frequent being phenylephrine (55%), norepinephrine (47%), ephedrine (31%), epinephrine (26%), and vasopressin (24%). Treatment was with anticoagulation in 88 (90%; therapeutic, 48%; prophylactic, 42%) and antiplatelet agents in 59 (60%; aspirin, 51%; clopidogrel, 15%). Other frequently associated conditions included mechanical ventilation at time of diagnosis (37%), diabetes (34%), peripheral arterial disease (32%), dialysis dependence (31%), cancer (24%), and sepsis (20%). Only five patients (5%) ultimately required finger amputation. The 30-day, 1-year, and 3-year survival was 84%, 69%, and 59%. By Cox proportional hazards modeling, cancer (hazard ratio, 2.4; 95% confidence interval, 1.1-5.6; P ¼ .035) was an independent predictor of mortality. There were 50 concurrent non-ICU patients with finger ischemia. Non-ICU patients were more likely to have connective tissue disorders (26% vs 13%; P ¼ .05) and hyperlipidemia (42% vs 24%; P ¼ .03) and to undergo finger amputations (16% vs 5%; P ¼ .03). Conclusions: Finger ischemia in the ICU is frequently associated with the presence of arterial lines and the use of vasopressor medications, of which phenylephrine and norepinephrine are most frequent. Anticoagulation or antiplatelet therapy is appropriate treatment. Whereas progression to amputation is rare, patients with finger ischemia in the ICU have a high rate of mortality, particularly in the presence of cancer. Non-ICU patients hospitalized with finger ischemia more frequently require finger amputations, probably because of more frequent connective tissue disorders. (J Vasc Surg 2018;68:1499-504.) Keywords: Finger; Vasospasm; Ischemia; Intensive care unit
Vascular surgeons may be consulted to evaluate patients in the intensive care unit (ICU) with digital ischemia. Several factors can lead to digital ischemia in these patients, including the severity of underlying illness From the Division of Vascular Surgery, Knight Cardiovascular Institute, Oregon Health & Science University. Author conflict of interest: none. Presented at the Thirty-second Annual Meeting of the Western Vascular Society, Blaine, Wash, September 23-26, 2017. Correspondence: Gregory J. Landry, MD, Division of Vascular Surgery, Knight Cardiovascular Institute, Oregon Health & Science University, 3181 SW Sam Jackson Park Rd, Mail Code OP11, Portland, OR 97239 (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.2018.01.050
(eg, sepsis,1,2 heart failure, malignant disease,3,4 renal failure,5 and diabetes6), use of vasoactive medications causing peripheral vasoconstriction,7-11 and use of radial arterial lines for continuous hemodynamic monitoring.12,13 Finger ischemia can be the source of great anxiety for critically ill patients, families, and care providers. Whereas these patients have multiple critical health issues by virtue of requiring ICU care, the hands are very visible, and finger abnormalities are a prominent outward sign of the patient’s illness. Although surgical intervention is not frequently required, vascular surgeons are often consulted to provide guidance as a vascular “specialist.” The immediate and long-term outcomes of patients in the ICU with finger ischemia are not well characterized and form the basis of this report. We sought to determine factors associated with digital ischemia in ICU patients and 1499
1500
Journal of Vascular Surgery
Landry et al
November 2018
to determine short- and long-term outcomes of the affected digits as well as short- and long-term survival of patients who develop finger ischemia in the ICU.
ARTICLE HIGHLIGHTS d
d
METHODS All patients in either medical or surgical ICUs at a single tertiary care facility (Oregon Health & Science University, Portland, Oregon) developing finger ischemia and who had vascular surgery referral from 2008 to 2015 were identified and evaluated. Patients were identified by querying the billing code for finger photoplethysmography (PPG) performed by an Intersocietal Accreditation Commission-approved vascular laboratory at this facility. The hospital records of all patients were evaluated, with data abstracted from the patients’ electronic medical record (Epic; Epic Systems Corporation, Madison, Wisc). Historical demographic data were recorded, including age, sex, body mass index, history of diabetes, hypertension, hyperlipidemia, peripheral arterial disease, chronic obstructive pulmonary disease, coronary artery disease, cancer, renal insufficiency or failure, and connective tissue disorders. We evaluated reason for ICU admission as well as additional concomitant ICU factors that could contribute to finger ischemia, such as the presence of radial arterial lines, mechanical ventilation, need for hemodialysis, and use of vasoactive medications. Type and duration of vasoactive medications were also recorded. Additional medications evaluated included anticoagulants, antiplatelet agents, antibiotics, beta blockers, and statin medications. Symptoms prompting referral and results of the finger PPG evaluation were also recorded. Active warming was used as needed to bring finger temperatures to at least room temperature. Abnormal findings included dampening or complete flattening of the PPG waveforms. Outcomes evaluated included the need for partial or complete finger amputation and overall survival. A concurrent control group of hospitalized, non-ICU patients undergoing finger PPG evaluation was also evaluated with the same information recorded. All categorical variables are presented as numbers and percentages. Continuous variables are presented as mean 6 standard deviation for normally distributed data and median with range for non-normally distributed data. Categorical data are compared with Fisher exact test or c2 test as appropriate. Continuous variables are evaluated with Student t-test. Kaplan-Meier analysis of survival was performed, and multivariate analysis of survival was performed using Cox proportional hazards modeling. The study was approved by the Institutional Review Board at Oregon Health & Science University. Because of the retrospective nature of the study and use of deidentified data, informed consent was not required for the enrolled patients.
d
Type of Research: Retrospective cohort study Take Home Message: In 98 intensive care unit patients with finger ischemia, the most frequent causes were the use of vasoactive medications and arterial lines. Finger amputations were needed in 5%, and finger ischemia was a marker for decreased survival, particularly so in patients with cancer. Recommendation: This study suggests that intensive care unit patients with finger ischemia have a low rate of finger amputation, but they have increased mortality, particularly if they have cancer.
RESULTS During the study period, 552 patients underwent finger PPG evaluations for upper extremity or digital ischemia. Of these, 404 were outpatient evaluations or duplicate studies and were excluded. Of the remaining patients, 98 were ICU patients and 50 were non-ICU inpatients. Of the 98 ICU patients, 55 were male and 43 were female. The mean age was 57.1 6 16.8 years. Of these patients, 42 were in the surgical ICU and 56 in the medical ICU. The symptoms were unilateral in 68 patients and bilateral in 30. Presenting symptoms included discoloration in all patients, with sensory findings (pain or paresthesias) in 47 (48%) and the presence of digital ulcerations or gangrene in 3 (3%). During the same period, 50 non-ICU patients with finger ischemia were identified to serve as a control group. Demographics of the ICU and non-ICU patients are included in Table I. In comparison to the ICU patients, non-ICU patients more frequently had hyperlipidemia (24 [25%] vs 21 [42%]; P ¼ .03) and connective tissue disorders (13 [13%] vs 13 [26%]; P ¼ .05), with a trend for increased diabetes (33 [34%] vs 23 [46%]; P ¼ .11) and dialysis (30 [31%] vs 21 [42%]; P ¼ .17). No other significant demographic differences were noted. Among ICU patients, findings on finger PPG studies were normal in 28 patients (29%). Unilateral abnormal findings were present in 40 (41%) patients, with the median number of fingers affected 3.5, mode 5. Bilateral abnormalities were noted in 30 patients (31%), with a median of 7.5 fingers affected, mode 10. Thus, when a hand is affected, it most frequently affected all digits in that hand. In comparison, non-ICU patients had a trend for more normal findings on PPG studies (n ¼ 21 [42%]) and unilateral PPG abnormalities (n ¼ 21 [42%]) and fewer bilateral abnormalities (n ¼ 8 [16%]; P ¼ .10). For each patient, the average digital pressure for all abnormal fingers was recorded. Among ICU patients, the digital pressures were lower than for non-ICU patients (54.7 6 45.9 vs 74.1 6 40.6; P ¼ .02). Correspondingly, digital brachial indices in ICU patients were also
Journal of Vascular Surgery
Landry et al
1501
Volume 68, Number 5
Table I. Demographic factors for intensive care unit (ICU) and non-ICU patients with finger ischemia Factor
ICU (n ¼ 98)
Non-ICU (n ¼ 50)
P value
Table II. Vasoactive medication frequency and duration in intensive care unit (ICU) patients with finger ischemia Patients (n ¼ 98), No. (%)
Medication
No. of days, mean 6 SD
Sex, M/F
55 (56)/43 (44) 24 (48)/26 (52)
.35
Norepinephrine
46 (47)
4.3 6 4.8
Age, years
57.1 6 16.8
56.6 6 17.4
.86
Phenylephrine
54 (55)
2.5 6 3.3
Smoking history
58 (59)
28 (56)
.78
Ephedrine
30 (31)
1.5 6 0.8
Hypertension
65 (66)
33 (66)
.97
Epinephrine
25 (26)
2.9 6 2.6
Diabetes
33 (34)
23 (46)
.11
Vasopressin
24 (24)
2.5 6 1.6
Peripheral vascular disease
31 (32)
19 (38)
.439
Dopamine
7 (7)
2.4 6 1.5
Dobutamine
6 (6)
2.3 6 1.0
Coronary artery disease
30 (31)
16 (33)
.80
SD, Standard deviation.
Congestive heart failure
19 (20)
8 (16)
.60
Hyperlipidemia
24 (25)
21 (42)
.03
Table III. Medical therapy for patients with finger ischemia
Cancer
23 (24)
10 (20)
.63
Medications
Dialysis
ICU (n ¼ 98), No. (%)
Non-ICU (n ¼ 50), No. (%)
P value
30 (31)
21 (42)
.17
Antibiotics
83 (85)
31 (62)
.002
Connective tissue disease
13 (13)
13 (26)
.05
Anticoagulation
88 (90)
33 (66)
<.001
Therapeutic
47 (48)
17 (34)
.19
Chronic obstructive pulmonary disease
12 (12)
6 (12)
.97
Prophylactic
41 (42)
16 (32)
.25 .80
Body mass index, kg/m2
27.4 6 9.9
26.7 6 10.0
.71
Categorical variables are presented as number (%). Continuous variables are presented as mean 6 standard deviation.
Antiplatelet
59 (60)
29 (58)
Aspirin
50 (51)
28 (56)
.57
15 (15)
7 (14)
.83
Beta blocker
71 (72)
29 (58)
.08
Statin
46 (47)
19 (38)
.30
Clopidogrel
ICU, Intensive care unit.
lower (0.45 6 0.36 vs 0.59 6 0.33; P ¼ .030). The lower values noted in ICU patients are likely to represent a higher number of patients with acute illnesses, resulting in completely flattened waveforms with undetectable digital pressures, most of which resolved with systemic improvement, in comparison to the non-ICU patients. Radial arterial lines were present in 36 patients (37%). Of those with arterial lines, the majority had unilateral symptoms (n ¼ 24 [66%]); however, 12 (33%) had bilateral symptoms. This was not statistically different from the distribution noted in patients without arterial lines. PPG results in patients with arterial lines were normal in 10 (28%), unilaterally abnormal in 14 (39%), and bilaterally abnormal in 12 (33%). Twenty-four patients with arterial lines underwent upper extremity arterial duplex ultrasound examination, with documented radial artery thrombosis in 16 (67%) but no thrombosis noted in 8 (33%). No finger amputations were required among patients with radial arterial lines. A number of concomitant ICU conditions were identified, including mechanical ventilation in 36 patients (37%), hemodialysis in 30 (31%), and septic shock in 19 (20%). Twelve patients (13%) had concomitant toe ischemia. However, the most frequent ICU-related factor was the use of intravenous vasoactive medications, present in 80 (82%) patients. The most frequent vasoactive medications and the duration of use are noted in Table II. Phenylephrine (n ¼ 54 [55%]) and
norepinephrine (n ¼ 46 [47%]) were the most frequently used medications. Norepinephrine was used for the longest duration (4.2 days), with the remaining medications used for 3 days or less. Most patients were taking more than one vasoactive agent, with the total number of vasoactive agents as follows: none in 18 patients (18%), one in 21 patients (21%), two in 23 patients (24%), three in 22 patients (22%), four in 11 patients (11%), and five in 3 patients (3%). Additional medications are listed in Table III. Antibiotic use was much more common in ICU patients than in non-ICU patients (85% vs 62%; P ¼ .02). Therapeutic and prophylactic anticoagulation, antiplatelet medications (aspirin or clopidogrel), beta blockers, and statins were commonly used in both ICU and non-ICU patients. Overall anticoagulation rates were higher in ICU patients (90% vs 66%; P < .001); however, there were no significant differences noted when anticoagulants were subdivided into therapeutic and prophylactic doses. Among patients treated with therapeutic anticoagulation, most were already being treated therapeutically for other thrombotic events or risk factors (74% ICU vs 71% non-ICU; P ¼ NS), with the remainder started on anticoagulation specifically for their hand ischemia. Among the ICU patients, only five went on to digital amputations. In comparison, amputations were performed in eight non-ICU patients (5% vs 16%; P ¼ .03).
1502
Journal of Vascular Surgery
Landry et al
November 2018
Fig 1. Survival of patients with finger ischemia by intensive care unit (ICU) status. S.E., Standard error.
Non-ICU patients more frequently had a history of prior finger amputations than ICU patients (6 [12%] vs 3 [3%]; P ¼ .06). ICU patients undergoing amputation underwent amputation of 1.6 6 0.6 fingers compared with 1.1 6 0.4 (P ¼ .08) in non-ICU patients. The ICU length of stay in ICU patients was 11.8 6 10.8 days. Overall hospital length of stay was 20.5 6 14.6 days, significantly greater than overall hospital length of stay in non-ICU patients (7.3 6 4.5 days; P < .001). Survival of ICU vs non-ICU patients is shown in Fig 1. The 30-day survival was significantly lower for the ICU patients (84% vs 96%; P ¼ .04). Survival at 1 year, 3 years, and 5 years after a finger ischemic event was 69%, 59%, and 51% in ICU patients and 79%, 63%, and 42% in non-ICU patients (P ¼ .798). Among ICU patients, multivariate analysis with Cox proportional hazards modeling revealed cancer as the only predictor of mortality (hazard ratio, 2.4; 95% confidence interval, 1.1-5.6; P ¼ .035). Among non-ICU patients, the only significant multivariate predictor of mortality was increasing age (hazard ratio, 1.06; 95% confidence interval, 1.00-1.11; P ¼ .023), with the mean age of 51.4 6 18.1 years for survivors and 63.2 6 14.4 years for nonsurvivors. Among patients with finger amputations, including both ICU and non-ICU patients, survival at 1 year, 3 years, and 5 years was 83%, 47%, and 17% compared with 75%, 66%, and 61% of those not undergoing amputation at the same time points (P ¼ .168; Fig 2). There were too few amputation events to stratify for ICU and non-ICU patients, although among the ICU patients, the longest survival of a patient undergoing a finger amputation was 21 months.
Fig 2. Survival of patients with finger ischemia by finger amputation status. S.E., Standard error.
DISCUSSION Patients in the ICU frequently have hemodynamic instability requiring the use of vasoactive medications. However, a potential adverse sequela of hemodynamic support is reduced end-organ perfusion. The extremities are particularly vulnerable, and several reports of upper and lower extremity digital malperfusion in patients receiving vasopressor agents have been published.7-11 Whereas bilateral involvement is most frequent, asymmetric unilateral involvement has been described as was seen in our series. Previous studies have shown pre-existing vascular disease, concomitant use of multiple vasopressor agents, and prolonged periods of hypotension to be risk factors for digital ischemia.10 Most reports in the medical literature have identified the association between norepinephrine and risk of digital ischemia.7-11 Norepinephrine was used in almost half the patients in our series and was the vasopressor agent used for the longest duration. Of the five ICU patients who required digital amputation, two were taking norepinephrine and one each was taking phenylephrine, vasopressin, and epinephrine. In a systematic review from the Scandinavian Society of Anaesthesiology and Intensive Care Medicine, norepinephrine was identified as the first-line vasopressor of choice, with moderate quality of evidence in patients with shock in general and septic shock and low-quality evidence for cardiogenic or hypovolemic shock.14 Whereas a maximal tolerable dose of norepinephrine has not been definitively established, it is known that a high dose and duration of norepinephrine are associated with a high peripheral vasoconstriction risk, and an optimal dose of 0.2 to 1.3 mg/kg/min is
Journal of Vascular Surgery
Landry et al
1503
Volume 68, Number 5
associated with a reduced risk of end-organ ischemia.7 Accidental arterial injections causing hand ischemia have been reported with other vasoactive agents, such as ephedrine15 and epinephrine.16 No intra-arterial injections occurred in the current series. Digital necrosis due to high-dose intravenous infusion of epinephrine and dopamine has been reported.17 The management of digital ischemia in ICU patients is not well established. Overall, prognosis is clearly related to the severity of underlying disease, and minimizing the dosage and duration of vasopressor use as much as possible is prudent. The use of anticoagulation or antiplatelet therapy is not well established, although these patients are often already taking these medications for other indications. In the current series, 60% of patients were receiving antiplatelet therapy and 90% either therapeutic or prophylactic anticoagulation at the time of diagnosis. In cases of presumed arterial thrombosis, these treatments are warranted; however, in cases of severe vasospasm, the addition of these agents is of questionable benefit. Finger and hand ischemia secondary to radial arterial lines has been previously described in multiple reports. Valentine et al13 described eight patients during a 5-year period. Five patients underwent surgical intervention for radial artery thrombosis, two were treated with anticoagulation, and one was treated conservatively. Rethrombosis after surgical repair with patch angioplasty was common, and all patients treated surgically required digital amputation, whereas no amputations were required in patients treated nonsurgically. Garg et al12 identified 30 patients with complications related to radial arterial lines, including 15 thrombotic episodes, 10 pseudoaneurysms, and 5 abscesses. All thrombotic cases underwent surgical thrombectomy, with three developing rethrombosis. No patients required digital amputations; however, in-hospital mortality of the cohort was 37%, reflecting the severity of illness of these patients. No amputations were required in our series; however, 30-day mortality was only 16%, probably reflecting the fact that our series included all patients with symptoms of digital ischemia, even if only mild discoloration, which may portend a more favorable prognosis than for those who require surgical intervention because of ischemia severity. Operations due to radial artery occlusion were rarely required. One patient underwent radial thrombectomy. One patient initially admitted for bilateral lower extremity ischemia due to thromboembolism subsequently developed right upper extremity ischemia due to axillary and brachial artery emboli requiring thromboembolectomy. One patient undergoing chemotherapy for lymphoma was found to have axillary and brachial artery occlusion in a pattern typical for temporal arteritis along with ulnar artery occlusion and underwent axillary to radial artery bypass. One patient admitted with
bacterial endocarditis developed left upper extremity ischemia while in the ICU due to a brachial artery embolus, for which she underwent embolectomy. The rarity of finger amputations in the ICU group probably reflects primarily a vasospastic etiology. Despite the occasional severity of presentation, the intense vasospasm brought on by the patient’s underlying critical illness in conjunction with vasoactive agents is typically reversible as the patient’s overall condition improves and vasoactive agents are withdrawn. With few exceptions, digital ischemia secondary to vasospasm portends a good prognosis with respect to tissue loss.18 In contrast, the higher rate of finger amputation in non-ICU patients is likely to reflect an association with fixed digital artery occlusion due to disorders such as chronic hemodialysis, diabetes, and connective tissue disorders, which are more strongly associated with subsequent tissue loss.5,18,19 The poor survival of patients with digital ischemia is noteworthy. Although it would be expected that ICU patients have decreased 30-day survival compared with non-ICU patients, mortality rates beyond 30 days were not different between ICU and non-ICU patients. Whereas the ICU patients had greater acuity of illness, non-ICU patients had a number of more frequent comorbidities, such as dialysis dependence, connective tissue disorders, hyperlipidemia, and diabetes, which can be associated with decreased long-term survival. This indicates that regardless of etiology, finger ischemia is a poor prognostic indicator for survival. The need for finger amputation was not associated with decreased survival, but this is likely due to the small number of patients in the amputation group and probably reflects a type II error. This study has several limitations. The patients were identified by billing codes for PPG. This study is routinely performed in all patients evaluated for finger ischemia in the ICU, so we believe this captures all of the relevant patients during this period. Nonetheless, it is possible that some patients with finger ischemia were not evaluated by the vascular surgery service and did not undergo PPG evaluation. Presumably, though, these would be patients with milder symptoms, as patients with more severe symptoms routinely receive vascular surgery consultation. Alternatively, these could be patients whose systemic illness was so severe and life expectancy so low that finger ischemia was not addressed. Our database includes only patients with manifestations of digital ischemia. Most ICU patients taking vasoactive agents do not develop digital ischemia and therefore are not included in our database. We therefore cannot determine the incidence of digital ischemia in patients taking vasoactive agents or comment on the relative safety of one vasoactive agent vs another with respect to this complication. Furthermore, the retrospective nature of the study made it difficult to determine the dosages of the vasoactive
1504
Journal of Vascular Surgery
Landry et al
November 2018
agents at the time of symptom manifestation because dosages were continuously adjusted in response to patients’ needs. We therefore are unable to determine whether there was a threshold dose that was more likely to lead to ischemic symptoms; however, this is a question that we plan to evaluate prospectively. The data regarding arterial lines are based on the presence of an arterial line at the time of consultation. It is possible that patients had a prior arterial line that was unknown to us and that the number of patients with arterial line-associated ischemia is underestimated.
6.
7.
8.
9.
CONCLUSIONS Although it is a source of anxiety for patients, families, and care providers, digital ischemia in ICU patients rarely leads to finger amputation, with finger recovery associated with patients’ recovery from their critical illness. Conservative therapy is indicated in most cases. Digital ischemia, however, is a poor prognostic sign, associated with high rates of both early and late mortality.
AUTHOR CONTRIBUTIONS Conception and design: GL, TL, EM, EJ, CA, AA, RM, GM Analysis and interpretation: GL Data collection: GL, CM, DA, BM Writing the article: GL, CM, DA, BM Critical revision of the article: GL, TL, EM, EJ, CA, AA, RM, GM Final approval of the article: GL, CM, DA, BM, TL, EM, EJ, CA, AA, RM, GM Statistical analysis: GL Obtained funding: Not applicable Overall responsibility: GL
REFERENCES 1. Johansen K, Hansen ST. Symmetrical peripheral gangrene (purpura fulminans) complicating pneumococcal sepsis. Am J Surg 1991;165:642-5. 2. Davis MD, Dy KM, Nelson S. Presentation and outcome of purpura fulminans associated with peripheral gangrene in 12 patients at Mayo Clinic. J Am Acad Dermatol 2007;57:944-56. 3. Chow SF, McKenna CH. Ovarian cancer and gangrene of the digits: case report and review of the literature. Mayo Clin Proc 1996;71:253-8. 4. Kopterides P, Tsavaris N, Tzioufas A, Pikazis D, Lazaris A. Digital gangrene and Raynaud’s phenomenon as complications of lung adenocarcinoma. Lancet Oncol 2004;5:549. 5. Yeager RA, Moneta GL, Edwards JM, Landry GJ, Taylor LM, McConnell DB, et al. Relationship of hemodialysis access to
10.
11.
12.
13. 14.
15.
16.
17.
18.
19.
finger gangrene in patients with end-stage renal disease. J Vasc Surg 2002;36:245-9. Ahmadi N, Hajsadeghi F, Gul K, Leibfried M, DeMoss D, Lee R, et al. Vascular function measured by fingertip thermal reactivity is impaired in patients with metabolic syndrome and diabetes mellitus. J Clin Hypertens 2009;11:678-84. Shin JY, Roh SG, Lee NH, Yang KM. Ischemic necrosis of upper lip, and all fingers and toes after norepinephrine use. J Craniofac Surg 2016;27:453-4. Simman R, Phavixay L. Bilateral toe necrosis resulting from norepinephrine bitartrate usage. Adv Skin Wound Care 2013;26:254-6. Chuang SS. Finger ischemia secondary to the synergistic agonist effect of norepinephrine and ergonovine and in a burn patient. Burns 2003;29:92-4. Daroca-Perez R, Carrascosa MF. Digital necrosis; a potential risk of high-dose norepinephrine. Ther Adv Drug Saf 2017;8: 259-61. McAllister SF, Cleary JD. Blue digits: cyanosis resulting from norepinephrine treatment. Ann Pharmacother 2005;39: 383-4. Garg K, Howell BW, Saltzberg SS, Berland TL, Mussa FF, Maldonado TS, et al. Open surgical management of complications from indwelling radial artery catheters. J Vasc Surg 2013;58:1325-30. Valentine RJ, Modrall JG, Clagett GP. Hand ischemia after radial artery cannulation. J Am Coll Surg 2005;201:18-22. Møller MH, Claudius C, Junttila E, Haney M, OscarssonTibblin A, Haavind A, et al. Scandinavian SSAI clinical practice guideline on choice of first-line vasopressor for patients with acute circulatory failure. Acta Anaesthesiol Scand 2016;60:1347-66. Kozar S, Kurnik G. Accidental intra arterial injection of ephedrine: what about the treatment with nitroglycerin? J Anesth Clin Res 2015;6:2. Velissariou I, Cottrell S, Berry K, Wilson B. Management of adrenaline (epinephrine) induced digital ischaemia in children after accidental injection from an EpiPen. Emerg Med J 2004;21:387-8. Reyes AJ, Ramcharan K, Harnarayan P, Mooteeram J. Symmetrical digital gangrene after a high dose intravenous infusion of epinephrine and dopamine following resuscitation from cardiac arrest. BMJ Case Rep 2016;2016. bcr2016217977. Landry GJ, Edwards JM, McLafferty RB, Taylor LM, Porter JM. Long-term outcome of Raynaud’s syndrome in a prospectively analyzed patient cohort. J Vasc Surg 1996;23:76-86. Landry GJ, McClary A, Liem TK, Mitchell EL, Azarbal AF, Moneta GL. Factors affecting healing and survival after finger amputations in patients with digital artery occlusive disease. Am J Surg 2013;205:566-70.
Submitted Nov 4, 2017; accepted Jan 22, 2018.