Why Not Life and Limb? Vasopressor Use in Intensive Care Unit Patients the Cause of Acute Limb Ischemia

Abstract

Background: Acute limb ischemia (ALI) of the upper extremity is a rare yet severe condition in intensive care unit (ICU) patients that generally leads to amputation. The aim of this study is to determine risk factors for development of upper extremity limb ischemia in ICU patients requiring vasopressor support. Methods: This is a retrospective study conducted from 2010 to 2015. Patients who received vasopressors during ICU admission were considered for the study. Patients were identified via Current Procedural Terminology (CPT) billing codes. ALI patients were matched to control patients based on diagnosis and Acute Physiology and Chronic Health Evaluation II score. Days on pressors, number of pressors, total doses, and level of ischemia were recorded. Primary end point was doses, types, and days on vasopressors. Secondary end point was level of ALI. Results: Patients in the ALI group were more likely to be started on a higher number of different types of pressors (2.6 vs 1.3 pressors). ALI patients received pressors for 8.5 days compared with 1.6 days in control patients, and received 12.8 doses compared with 3.0 doses in control patients. In addition, vasopressors with alpha-adrenergic activity were more likely to be used in the ALI group. Level of ischemia was not linked to any of the tested variables. Conclusion: Patients admitted to the ICU are more likely to sustain an acute ischemic event of an upper extremity with more vasopressor usage. Patients who received alpha-adrenergic activating vasopressors were more likely to sustain limb ischemia. When discoloration of an extremity is detected, patients should receive counteractive treatments in an effort to salvage the extremity and prevent function loss.

Introduction

Acute limb ischemia (ALI) of the upper extremity is a rare yet severe condition that may lead to amputation of the fingers and/or hand. The condition is particularly recognized to occur in critically ill intensive care unit (ICU) patients.⁶ Necrosis is primarily caused by diminished or total loss of blood supply to body tissues. The compromised blood supply may result from trauma, infection, chronic vascular insufficiency, and iatrogenic causes.^6,15,18 Other risk factors include diabetes mellitus, long-term smoking, connective tissue disorders, malignancies, vasculitis, thromboembolism, and renal diseases.^17,19

Blood pressure support in critically ill patients may be a necessary life-preserving treatment in hypotensive patients. Ionotropic and adrenergic medications are commonly used to maintain perfusion pressure to vital organs. However, prolonged use of vasopressors may be associated with ischemia of the digits in the upper and lower extremities.^1,13,14 Often, the ischemic effects are irreversible, resulting in life-altering amputations. The prevalence and incidence of digital necrosis of the upper extremity are difficult to establish because it is usually a secondary diagnosis upon hospital admission and may not be accurately recorded in patients’ hospital records. In addition, referral to a hand surgeon is often delayed for several days after the onset of acrocyanosis. In many cases, no early treatment is initiated despite progressive tissue ischemia.

To date, there is a paucity of literature regarding upper extremity digital necrosis in critically ill patients requiring vasopressor support. Oftentimes, it is seen as a tolerable outcome in the critically ill patient due to the old medical adage “life over limb.” Unfortunately, in practice digital necrosis is a potentially limb-threatening condition, which may have a significant impact on normal function and independence.¹⁶ Short-term survivability in patients with ALI is poor, with 30-day amputation rates from 10% to 30% and mortality rates of 15%.⁷ The aim of this study is to determine the risk factors for the development of upper extremity digital necrosis in ICU patients requiring vasopressor support.

Methods

A retrospective chart review was conducted from 2010 to 2015 at our tertiary care center with 7 separate ICUs. Institutional review board approval was sought through our institution and exemption was granted due to retrospective nature of study. Patients who received vasopressors during their ICU admission were considered for the study. Patients with ALI were identified through the orthopedic department patient tracking system, which provides information on all past inpatient consultations. Control patients were identified via Current Procedural Terminology (CPT) billing codes for vasopressor use. Control patients were those who received vasopressors but did not get ALI. ALI patients were matched to control patients based on reason for ICU admission and Acute Physiology and Chronic Health Evaluation II (APACHE II) score initially at a 1:2 ratio (Table 1). The attending hand surgeon on call made determination of digital limb ischemia at the time of consultation. Exclusion criteria included death prior to average days to ischemia time, nonmatched reason for ICU admission, incomplete or unavailable electronic medical record (EMR), and lack of vasopressor use during ICU stay. Level of ischemia was assigned a numerical value for tracking purposes.

Table 1.

Summary of Patient Diagnosis and APACHE II Score.

Acute limb ischemia patients			Diagnosis	Matched control patients
Patient	ICU days	APACHE II		Patient	ICU days	APACHE II
1	15	22	Great vessel injury	1	1	13
				2	24	28
2	10	21	Respiratory failure	3	1	26
				4	1	11
3	16	17	Aortic dissection	5	1	34
				6	6	22
4	117	20	Mitral valve disorder	7	2	26
				8	33	16
5	26	16	Sepsis	9	1	11
				10	1	15
6	64	23	NSTEMI	11	3	5
				12	7	6
7	29	28	Interstitial lung disease	13	9	24
				14	56	31
8	58	15	Interstitial lung disease	15	43	11
				16	77	12
9	12	30	Alveolar hemorrhage	17	2	26
				18	2	9
10	36	29	COPD exacerbation	19	7	29
				20	45	13
11	4	33	Sepsis	21	1	39
				22	8	32
12	28	42	Sepsis	23	12	17
				24	14	35
13	8	15	Sepsis	25	3	23
				26	10	17
14	26	25	Cardiac arrest	27	1	25
				28	4	18
15	11	23	Sepsis	29	1	36
				30	31	28
16	5	24	Attempted suicide	31	14	19
				32	57	39
17	33	37	Sepsis	33	3	39
				34	14	37
18	29	32	Stroke	35	12	22
				36	2	19
19	8	38	Endocarditis	37	3	37
				38	4	33
20	6	32	Cardiogenic shock	39	1	25
				40	2	24
21	23	18	Small bowel obstruction	41	79	23
				42	54	19
22	9	37	Pulseless electrical activity	43	1	32
				44	3	9
23	3	13	Aortic stenosis	45	3	11
				46	7	9

Note. Summary table of all included patients in the study, with cross-matching based on admission diagnosis and APACHE II score performed. Those in gray were removed due to death prior to average time to ischemia. ICU = intensive care unit; COPD = chronic obstructive pulmonary disease; APACHE II = Acute Physiology and Chronic Health Evaluation II; NSTEMI = non-ST-elevation myocardial infarction.

Following application of exclusion criteria, there were 23 ALI patients and 36 cross-controlled patients for final analysis. Average time to ischemia in the ALI patients was 8.65 days; therefore, 10 cross-matched control patients were excluded from final data analysis due to death occurring prior to average ischemia time (<9 days). Figure 1 provides a detailed breakdown of patient enrollment. All patient data were recorded from time to admission in ICU to time of either death or discharge out of the ICU. Patients at our hospital cannot be receiving vasopressor support outside of the ICU; therefore, time in the ICU was able to catch all data points. First day of ischemia was deemed to be the date at which limb color was questioned to be concerning. Identifying phrases searched for in the chart included, but was not limited to, “dusky,” “discolored,” “ischemic,” “cool to touch,” “ecchymosis,” “loss of circulation,” “digital edema,” and “gangrene.” If no mention in the chart, the date of orthopedic consult was used as the first day of ischemia. Of note, all patients whom orthopedic consultation was used as the first day of ischemia had early-stage ischemia without irreversible changes. Vasopressors used in our ICU, for the purpose of this study, include dopamine, dobutamine, epinephrine (EPI), norepinephrine (NE), and phenylepinephrine (PhE). Standard doses of the medications used in this study are as follows: dopamine (400 mg in D5W 500 mL), dobutamine (500 mg/250 mL intravenous piggyback, EPI (1 mg of 1 mg/10 mL solution), NE (4 mg), PhE (80 mg of 1% 10 mg/1 mL solution). Information gathered included reason for admission to the ICU; comorbidity severity (calculated via the APACHE II score); days of ICU stay; type, number, and total dosage of vasopressors given; number of days on vasopressors; level of ALI; and treatment of the ischemia, among other data. Primary end point of the study was number of vasopressor doses, types of vasopressors used, and days on vasopressor support. Secondary end point was level of ALI.

Figure 1.

Flowchart showing the recruitment process and how patients were excluded; also demonstrates the cross-matching of admission diagnosis between the ischemia and control groups.

Statistical analysis was undertaken using both parametric (t test and analysis of variance) and nonparametric (Wilcoxon and Kruskal-Wallis) testing. Fisher exact test analysis was used for types of vasopressors used and level of ischemia. Nonparametric analysis was used for number of days on vasopressor support and total doses of vasopressors. Statistical significance was defined as a P value < .05.

Results

Demographic data between the 2 groups are summarized in Table 2. There was no statistical significance between age, sex, and number of days in the ICU. Patients in the ALI group were more likely to be started on a higher number of different types of pressors (P < .001). The ALI patients received on average 2.6 types of vasopressors compared with 1.3 in the control patients (P < .0001). No patient was put on all 5 vasopressors during ICU stay. In addition, patients who developed ALI had more days on vasopressor support and received more total doses of vasopressors. ALI patients on average received pressors for 8.5 days compared with 1.6 days in the control patients (P < .0001), and received on average 12.8 doses of vasopressors compared with 3.0 doses in the control patients (P < .0001). In addition, it was discovered that vasopressors with alpha-adrenergic activity were more likely to be used in the acute ischemic patient group, with 69.6% receiving EPI, 91.3% receiving NE, and 52.8% receiving PhE compared with 33.3% EPI (P = .006), 41.7% NE (P < .0001), and 22.2% PhE (P < .0001) in the control groups. Results are summarized in Table 3 and 4.

Table 2.

Demographic Data of Patients.

	Age	Sex	ICU days	Death
Acute ischemia	60.96	14 Male 9 Female	25.04	11/23
Control	61.03	19 Male 17 Female	17.89	3/36
P value	.983	.548	.281	.002

Note. Comparative demographic data for ALI vs non-ALI patients during their ICU stay. ALI = acute limb ischemia; ICU = intensive care unit. Bold-faced value is statistically significant.

Table 3.

Different Types of Vasopressors Given.

Attribute	Ischemia	Control	Total	P value	Test method
Total No. of pressors, n (%)				<.0001	Fisher exact
1	2 (6.3%)	30 (93.8%)	32 (100.0%)
2	10 (71.4%)	4 (28.6%)	14 (100.0%)
3	7 (87.5%)	1 (12.5%)	8 (100.0%)
4	4 (80.0%)	1 (20.0%)	5 (100.0%)
Total	23 (39.0%)	36 (61.0%)	59 (100.0%)

Note. Representation of the total number of types of vasopressors patients in the ALI vs control group were on during their ICU stay. Bold-faced value is statistically significant. ALI = acute limb ischemia.

Table 4.

Days and Types of Vasopressors.

	No. of VP	Days on VP	No. of doses	Dopamine	Dobutamine	EPI	NE	PhE
Ischemia	2.6	8.5	12.8	39.1%	4.3%	69.6%	91.3%	52.8%
Control	1.3	1.6	3.0	22.2%	8.33%	33.3%	41.7%	22.2%
P value	<.0001	<.0001	<.0001	.168	.560	.006	<.0001	<.0001

Note. Representation of total vasopressor doses, number of days on vasopressors, and types of vasopressors patients received during their ICU stay. Bold-faced values are statistically significant. VP = vasopressor; EPI = epinephrine; NE = norepinephrine; PhE = phenylepinephrine.

Finally, the number of days on vasopressor support, the total doses of vasopressors used, and the types of vasopressors used had no correlation to the level of ischemia that occurred. Patients were not more likely to have a higher level of ischemia if more vasopressor types were used (P = .9171). In addition, when evaluating number of doses, there was no correlation to a higher level of ischemia (P = .8878), and when evaluating the number of days on vasopressor support, there was no correlation to a higher level of ischemia (P = .8889).

Discussion

To better understand the clinical effects of pressure-supporting medications, we must first review the mechanisms of action of the different drugs in our study. In general, pressure-supporting medications are broken down into inotropes or vasopressors (both commonly referred to as vasopressors in this study). Inotropes act centrally to increase myocardial contractility, whereas vasopressor agents are administered to increase vascular tone.³ The medications in our study act on largely the dopamine, alpha-adrenergic, and beta-adrenergic receptors of the body. Subcategory breakdown finds that alpha-1 & -2 and beta-1 & -2-receptors are found in the arterial system and result in vasoconstriction by increasing calcium.² It is important to note that while both receptor types are found in the major arterial systems (alpha-receptors in large arteries only, beta-receptors in small>large arterial systems), the cutaneous circulation contains only alpha-2-receptors.^5,11 Therefore, when alpha-2-receptors are preferentially activated via EPI, NE, and PhE, the cutaneous circulation is more likely to suffer. In addition, these same alpha-receptors have never been identified in myocardium tissue and therefore are unlikely to have a central impact on cardiac output.^3,5 Table 5 provides a complete breakdown of mechanisms of action of the vasopressors used in our study. These mechanisms of action fall in line with our results, which show that patients who had ALI were more likely to have received alpha-adrenergic receptor medications than those who did not.

Table 5.

Mechanism of Action of Vasopressors.

Agent	Alpha 1	Beta 1	Beta 2	Dopamine
Dopamine	+ +	+ + + +	+ +	+ + + + +
Dobutamine	+	+ + + + +	+ + +	0
EPI	+ + + + +	+ + + +	+ + +	0
NE	+ + + + +	+ + +	+ +	0
PhE	+ + + + +	0	0	0

Note. Demonstration of the receptor activity of each vasopressor assessed on our study. + = level of activity on receptor (ie, more +, more activity). EPI = epinephrine; NE = norepinephrine; PhE = phenylepinephrine.

In current septicemia protocols, the initiation of peripheral vasoactive medications for pressure support precedes the use of centrally acting inotropic medications.¹² Namely, vasoactive agents are to be used for a mean arterial pressure (MAP) <65 mm Hg, with inotropes being introduced only once MAP remains <70 mm Hg with the initiation of vasoconstricting medications.¹² This protocol has been demonstrated time and again to improve septicemia outcomes and survivability.^10,12,21 Unfortunately, commonly in our patient base, no matter the cause for admission to ICU, septicemia protocols using vasoactive agents first would be applied to all-comers. A study by Gurnani et al has shown that adequate fluid resuscitation in ICU patients, particularly those with septicemia, has shown to decrease the requirement for vasopressor support.¹⁰ Indeed, after the initiation of strict septicemia protocols upon the adoption of an electronic ordering system 1 year ago with particular MAP criterion having to be met prior to the administration of vasopressor support, our ALI rates have decreased (odds ratio = 3.1; P = .11; number needed to treat [NNT] = 101). Our institution’s protocol calls for initial fluid resuscitation x 3 hours. After 3 hours, if MAP is <65 mmHg then vasoconstrictive pressors should be started. If MAP is >65 mm Hg, then intravenous (IV) resuscitation should continue. After 1 hour of further IV resuscitation, if patient remains hypotensive (systolic <90 mmHg), then vasopressor agents should be started. Low cardiac output should be treated with dobutamine and blood transfusion. This further supports the idea of overly aggressive vasopressor use increasing the incidence of ALI in patients who may be better suited to fluid resuscitation. Due to this discovery, it is our recommendation that if possible, vasopressor use should be discontinued as soon as possible and, if not feasible, changed away from alpha-adrenergic activators when possible.

In our practice, we are oftentimes called early in the acute ischemia phase, where digits are often described as “dusky” without clear evidence of irreversible necrosis. The determination whether ischemia is reversible is somewhat subjective, but largely based on appearance of soft tissue and amount of necrotic tissue. Initial management at our institution recommends the use of nitrogen paste to the affected fingers at first presentation of abnormal color. Paste should be applied in a thin layer covering the skin that is abnormal in color. In addition, we recommend the use of calcium channel blockers to assist with peripheral vasodilation, when possible, from a hemodynamic standpoint. Finally, temperature regulation is recommended with indirect heat or warming blankets being used on the affected limbs. At this institution, anticoagulation is administered on all ICU patients unless a major contraindication is present. Therefore, while our algorithm does not include anticoagulation initiation, such consideration should be made if not already being administered. If consultation to the hand surgery department is made at a point where ischemia is not reversible, recommendations alter. In this situation, if the patient remains on vasopressors in the ICU, we await the discontinuation of vasopressor support and hemodynamic stabilization before performing initial amputation. Awaiting the full demarcation allows the patient to undergo a single, definitive operation with better healing results due to healthy tissue being closed. If the patient has stabilized from a hemodynamic standpoint and is removed from vasopressor support, we will offer amputation at that time. If the patient’s necrosis is contributing to the overall requirement of ICU pressure support (ie “wet” gangrene), a temporizing amputation will be performed with definitive amputation to occur once stabilized. For complete treatment algorithm, refer to Figure 2.

Figure 2.

Decision tree demonstrating the treatment algorithm upon initial consultation of suspected acute limb ischemia.

Note. Note the early differentiation of reversible vs nonreversible ischemia, which may not be able to be defined until conservative management has failed. It is important to consistently reassess the patient’s clinical status as they are maintained on vasopressors to account for any progression of ischemia.

In addition to the above treatment algorithm, research into other options to treat finger ischemia has been performed. Botox injections have long been used as a treatment option for Raynaud’s phenomenon,^9,20 with improvement in pain, oxygen saturation, and blood flow to affected digits.^9,20 Despite this, no quality studies exist on the efficacy of botox injections on pharmacologically induced digital ischemia, and therefore treatment algorithm recommendations are difficult. Botox injections should be considered if the patient is unable to be weaned from vasopressor support in the near future.

Sympathectomy has been suggested to be beneficial in limiting ischemic injury.^8,22 A pilot study evaluated thoracoscopic sympathectomy in limiting digital ischemia experienced by drug abusers after intra-arterial injection. They were able to prevent amputation in 81% of their patients using this technique, though best outcomes were achieved in those with minimal presenting ischemia.⁸ Finally, a study evaluating axillary block-induced chemical sympathectomy in digital ischemia patients found improved blood flow to the radial and ulnar arteries at the wrist 1 hour after injection.²² The study did state though that benefits generally returned to baseline at 1 week, and their results were better suited to rheumatologic conditions than anatomic or atherosclerotic etiologies.²² Again, no studies specifically evaluate sympathectomy use in pharmacologically induced ischemia. A role for sympathectomy in vasopressor-induced ischemia will likely require an indwelling catheter to deliver continuous blockage and as such will require a risk evaluation in patients who may already be suffering from septicemia. Such a treatment option should be considered if the patient is unable to be weaned from vasopressor support in the near future.

Limitation of this study include its retrospective nature of data collection limiting how much can be extrapolated from the chart. We can be sure that all patients were gathered via our tracking system, as at our hospital no other department (including vascular surgery, plastic surgery, or general surgery) manages hand injuries of any kind. First day of identified ischemia was subjective due to the retrospective data collection, and therefore we may have over- or underestimated the average time to ischemia. As such, control patients may have erroneously been excluded based on the average time to ischemia. In addition, the statistically significant higher rate of death in our ALI patients may lend itself to the argument that these patients were sicker overall. Finally, though trying to normalize the data through APACHE II scores, we were unable to completely eliminate confounding comorbidities, which may contribute to overall rates of ALI in our patient population. An analysis of all possible confounding comorbidities is not feasible in this retrospective study, but APACHE II scores generally serve as a good substitute for overall medical health, with higher scores trending toward patients with higher numbers of comorbidities.⁴

Conclusion

Patients admitted to the ICU who were found to have an ischemic event of the extremity remained on vasopressors for a longer period of time, received more doses of vasopressors, and received multiple types of vasopressors. Patients’ who received alpha-adrenergic activating vasopressors were more likely to sustain limb ischemia than those who received beta-adrenergic or dopamine-activating vasopressors. It is our recommendation that vasopressor use should be tempered with appropriate fluid resuscitation in the setting of ICU admission. In addition, when discoloration of an extremity is first detected, patients should receive immediate counteractive treatments, and attempts should be made to discontinue alpha-adrenergic vasopressors in an effort to salvage the extremity and prevent further function loss.