Acute limb ischemia is a rare complication of deep venous thrombosis (DVT) . Phlegmasia cerulea dolens (PCD) and the subsequent development of venous gangrene are rare ischemic events that result from acute massive venous thrombosis of the total or near-total venous outflow of an extremity. Although the incidence of PCD and venous gangrene is unknown, the incidence of all forms of venous thromboembolism (VTE) has been estimated to be as high as one million cases per year in the United States [1]. Upper extremity PCD is rare, compared to lower extremity PCD, occurring in 2–5 % of all reported cases of PCD [2, 3].

PCD patients present with ischemic pain, cyanotic skin discoloration, and massive extremity swelling. Venous outflow obstruction causes massive fluid sequestration in the extremity, leading to shock with resulting mortality rates of 25–40 % [4]. The ischemia with PCD is secondary to venous hypertension and is reversible if treated early and aggressively. However, PCD may progress to irreversible venous gangrene in 40–60 % of patients, resulting in amputation rates of 12–50 % [4–7]. Patients with PCD also have a higher incidence of pulmonary emboli (12–40 %) than patients with submassive deep venous thrombosis, while 30 % of deaths reported from PCD are due to pulmonary embolism [3, 6, 8]. Successful management of both lower and upper extremity PCD is dependent upon early diagnosis, anticoagulation, aggressive fluid resuscitation, and relief of venous hypertension.

The recognition and understanding of PCD as a clinical entity stretches back to the sixteenth century when, according to Haimovici, Fabricius Hildanus first proposed in 1593 that extremity gangrene could be secondary to venous thrombosis [9]. Hueter, in 1859, was the first to clearly outline the pathologic and clinical criteria for the diagnosis of “gangrene of venous origin” [5]. In 1862, Cruveilhier described accurately that venous thrombosis in these patients involved simultaneous thrombosis of the deep and superficial veins of an extremity [7]. Sixty-two years later, in 1924, Buerger described that gangrene can occur with venous thrombosis without larger artery thrombosis [5]. Fontaine and deSouza-Pereira demonstrated in 1937 that total venous occlusion of the lower extremity in a canine model resulted in gangrene [7]. The next year, Gregoire published a detailed description of ischemic venous thrombosis and used the term phlegmasia cerulea dolens (painful, blue, inflammation) to differentiate the condition from the more common nonischemic presentation of iliovenous thrombosis, termed phlegmasia alba dolens (painful, white, inflammation) [5]. Gregoire hypothesized that arterial spasm was responsible for the development of ischemia; however, his own use of acetylcholine and periarterial sympathectomy failed to achieve limb salvage. Amputation prevented mortality. Other authors in the 1930s, including Leriche, attributed the development of edema and ischemia to peripheral vasospasm [7]. DeBakey and Ochsner, in 1939, found the presence of “arterial spasm” after chemical phlebitis or iliofemoral vein ligation which was relieved by sympathectomy [7]. However by 1949, when they reported the largest series (n = 56) of PCD patients published in an American journal to that time, DeBakey and Ochsner attributed the development of ischemia to venous obstruction, stating that vasospasm was a less plausible explanation [10]. Brockman and Vasco created a canine hind limb model of ischemic venous thrombosis consisting of ligation of the entire iliofemoral venous system with the filling of the distal veins with inert barium sulfate [8]. The dogs exhibited massive extremity swelling, cyanotic mottling of the skin, loss of distal pulses, tachycardia, and shock, with death ensuing in less than 24 h. No change in anterior tibial arterial pressures or flow amplitudes occurred after venous ligation, indicating no primary role for vasospasm in the development of PCD. The authors concluded that the cause of ischemia with venous occlusion is the collapse of arterioles and small arteries due to the combination of marked increases in tissue pressure, secondary to edema, and a drop in hydrostatic pressure due to shock. While the exact mechanism of ischemia remains unclear, Brockman and Vasco’s conclusions are shared by most practitioners today.

Venous obstruction causes changes in the normal homeostasis between the influx and outflow of fluid occurring at the venous and arterial capillary beds. These changes lead to massive sequestration of fluid in the interstitial space, shock, collapse of arterioles and small arteries, and eventually ischemia and gangrene. Normal arterial capillary pressure exceeds the colloid osmotic pressure by 6–7 mmHg, leading to the passage of fluid into the interstitial space. The tissue pressure outside the capillaries is normally near zero. The colloid osmotic pressure at the venous end of the capillary normally exceeds the venous hydrostatic pressure by 11–15 mmHg which forces fluid back into the venous end of the capillary [8]. Massive venous occlusion causes the venous hydrostatic pressure to exceed the osmotic pressure, preventing absorption of fluid at the venous end of the capillary resulting in accumulation of interstitial edema. Venous pressure may increase 16–17-fold within 6 h of occlusion [11]. As the edema progresses, failure of extremity lymphatic outflow compounds the problem. Fluid continues to accumulate until a drop in hydrostatic pressure occurs due to arterial collapse. Interstitial pressures can reach 21–48 mmHg within 48 h after occlusion [8, 11]. Abnormal resting intramuscular compartment pressures in patients with PCD have been measured in the range of 37–70 mmHg [12–14]. The development of acute compartment syndrome with its coincident decrease in arterial perfusion acts as a second cause of ischemia in the extremity. Fasciotomy may improve limb salvage rates in patients with PCD and venous gangrene who develop acute compartment syndrome when combined with anticoagulation; however, fasciotomy is likely most effective if combined with attempts to relieve the venous obstruction [13–15].

Clinical deterioration to a state of shock occurs in patients with PCD due to massive sequestration of fluid in the extremity with venous obstruction. In a canine model of PCD, Brockman and Vasco reported loss of half of the blood volume into the extremity over 8 h. Intravenous volume replacement led to further third spacing of fluid and increased extremity edema. Haller estimated that 6–10 L of edema fluid builds up in a patient’s extremity after 3–5 days of venous occlusion [16]. Therefore, rapid, aggressive fluid resuscitation is a critical element of the treatment of PCD.

All the factors contributing to the development of ischemia in patients with PCD are not known with clarity. The generally accepted theory of why ischemia develops in PCD involves a “critical closing pressure ” for arteries that once exceeded leads to arterial collapse and subsequent arterial insufficiency and gangrene. Burton proposed that there are two forces at play in the arterial wall, transmural pressure and tension, that when in equilibrium, keep the vessel open [17, 18]. Hydrostatic pressure acts to distend the vessel, while wall tension acts to decrease the vessel’s radius. Interstitial pressure normally approaches zero. The transmural pressure is the difference in venous hydrostatic and interstitial pressures. As interstitial pressure rises or hydrostatic pressure drops, the transmural pressure decreases. With venous obstruction, the interstitial and venous hydrostatic pressures initially markedly increase. With the onset of shock, the hydrostatic pressure drops while the interstitial pressure continues to rise. Eventually the transmural pressure will drop to a “critical closing pressure” that is exceeded by the wall tension, leading to vessel collapse [8, 18]. Brockman observed this phenomena in his canine model of PCD and demonstrated that vasospasm did not contribute to ischemia as the hemodynamics were the same after sympathectomy [8].

All known risk factors for the development of venous thromboembolism are also associated with phlegmasia cerulean dolens and venous gangrene [18]. Reviews published in the 1960s by Brockman and Vasco and Haimovici found the most common risk factors associated with PCD to be: postoperative state (14–22 %), malignancy (12–18 %), postpartum condition (8–15 %), trauma (6–8 %), history of venous thromboembolism (3–11 %), inflammatory bowel disease (4–5 %), heart failure (5 %), and unknown (9–18 %) [3, 5]. Haimovici noted that patients with venous gangrene had a higher association with malignancy than patients with PCD but without gangrene (25 % vs. 12 %, respectively) [3]. A modern review of 62 patients with PCD identified between 1980 and 2008 found that malignancy is the strongest risk factor (33.9 %) today associated with the development of PCD and venous gangrene (Table 28.1) [4].

Trousseau in 1865 observed the association between malignancy and venous thrombosis when he described recurrent migratory thrombophlebitis in patients with cancer [19]. Patients with cancer have been estimated to have a 4–20 % chance of experiencing venous thrombosis [20]. Abnormalities of blood coagulation have been reported in up to 92 % of cancer patients [19]. Patients have presented with PCD suffering from numerous types of cancer including pancreatic carcinoma, bronchoadenocarcinoma, testicular cancer, thyroid cancer, renal cell carcinoma, cholangiocarcinoma, and ovarian adenocarcinoma [21]. Between 10 and 16 % of patients with PCD will present without a discernable etiology, but 10 % of them will be diagnosed with a malignancy within 1 year [4, 22, 23]. While the exact pathogenic mechanism for the development of venous thrombosis in cancer patients remains elusive, we know it involves complex interactions between the tumor cell, host cells, and the coagulation system [23]. Tumors activate blood coagulation in part by abnormal expression of high levels of tissue factor (TF) . Tissue factor is a constitutively expressed procoagulant, which causes the activation of the patients’ extrinsic clotting pathway [24]. In addition to TF, plasma levels of factor VIIa, factor XIIa, and the thrombin–antithrombin complex are elevated in patients with malignancy [25]. Patients with PCD are more likely to progress to venous gangrene if they have cancer [3].

Patients with a hereditary hypercoagulable state (14.5 %; Table 28.1) constitute the second largest group of patients at risk for developing PCD and venous gangrene [4]. Patients with PCD have been diagnosed with factor V Leiden and prothrombin 20210A gene mutations and deficiencies in plasminogen, proteins C and S, and antithrombin [22]. Antiphospholipid syndrome (APS), an acquired thrombophilia, is also associated with the development of PCD. Diagnosis is confirmed after a thrombotic event with the finding of lupus anticoagulant or either anticardiolipin or beta-2-glycoprotein I antibodies [21]. Heparin-induced thrombocytopenia (HIT) may lead to PCD and venous gangrene [26]. Venous thrombosis as a complication of HIT occurs four times more frequently than arterial thrombosis [27]. Warfarin treatment of HIT-associated deep venous thrombosis will lead to progression to venous gangrene in approximately 10 % of patients [27]. The development of venous gangrene in HIT patients with DVT is thought to be secondary to the warfarin-induced failure of the protein C anticoagulant pathway to downregulate the increased thrombin generation that occurs with HIT.

PCD is known to develop in patients with left common iliac vein compression by the right common iliac artery (May–Thurner syndrome) and in patients with inferior vena cava (IVC) filters [28, 29]. PCD has also been reported to occur in patients who develop DVT around an indwelling femoral venous catheter [30]. In Chinsakchai et al.’s recent review, the combined frequency of venous compression or venous indwelling foreign bodies was 6.4 % in patients with PCD [4]. Other causes of PCD and venous gangrene include immobilization, recent trauma, age, ulcerative colitis, mitral valve stenosis, intravenous drug use, venous stasis, the use of pharmacologic contraception, and pregnancy [4, 15, 18, 21, 22, 26, 28, 31].

PCD is estimated to occur in 2–5 % of patients with upper extremity deep vein thrombosis (UEDVT) , while UEDVT accounts for up to 4 % of documented cases of DVT [2, 3, 32]. The etiology of UEDVT can be divided into primary and secondary causes. Primary UEDVT includes cases of effort-induced thrombosis (Paget–Schroetter syndrome ) as well as unprovoked UEDVT. Secondary UEDVT accounts for 80 % of all cases [14]. Predisposing secondary factors include central and peripherally inserted central venous catheters (48 %), malignancy (38 %), immobility (14 %), prior DVT (7 %), implanted pacemakers, oral contraceptives, and left ventricular heart failure [14, 33]. Patients with UEDVT with an associated malignancy, hypercoagulable condition, or low cardiac output states are most likely to progress to PCD and/or venous gangrene [14].

Venous thrombosis can be viewed as a disorder with a spectrum of clinical presentations. At its most benign, venous thrombosis presents with painless calf swelling due to calf vein thrombosis. As the thrombus burden extends into the proximal outflow veins of an extremity, the edema becomes more pronounced, the skin is pale, and the limb may be painful. The skin is pale and blanches without cyanosis due to subcutaneous edema without venous congestion. This stage of deep venous thrombosis has been termed phlegmasia alba dolens and is a common clinical presentation of iliofemoral venous thrombosis [22]. Presentation with phlegmasia alba dolens was termed “milk leg” in the past due the preponderance of pregnant and postpartum woman with the condition presenting with iliac vein thrombosis secondary to uterine compression of the left common iliac vein against the pelvic rim. Patients presenting with phlegmasia alba dolens have significant swelling secondary to outflow obstruction, but they do not suffer from cyanosis and limb ischemia as they have patent venous collaterals. Without adequate therapy, patients with phlegmasia alba dolens will progress to phlegmasia cerulea dolens following the thrombosis of their venous collateral pathways. PCD is preceded by phlegmasia alba dolens in 50–60 % of patients [3, 15, 22].

The clinical presentation of patients with phlegmasia cerulea dolens is the triad of massive extremity swelling, cyanotic to purple skin discoloration, and severe ischemic extremity pain (Fig. 28.1). Although the onset of symptoms may be gradual, the majority of patients present acutely with progressive extremity swelling that takes on a tense and firm quality that is often described as “woody” in character but may be pitting. Ongoing fluid sequestration leads to the development of cutaneous blebs and bullae. Skin cyanosis progresses to mottling and progressive blue/purple skin discoloration. The changes start distally and progress proximally at varying rates. Bullae and cyanotic changes are most intense in the feet. Pain, due to ischemia, is a constant feature and is often agonizing and difficult to manage successfully. Ischemia is due to the venous obstruction coupled with the massive swelling and in some cases the development of acute compartment syndrome. Pain begins in the calf or femoral triangle but soon affects the entire extremity. Arterial insufficiency is progressive in untreated PCD. Palpable pedal pulses are present only in 17 % of patients with PCD, likely due in part to massive pedal edema [5]. Arterial continuous-wave Doppler signals are usually intact in uncomplicated PCD; however, loss of signals heralds venous gangrene (Table 28.2). The development of sensory and motor impairment occurs as the severity of PCD progresses from uncomplicated to impending gangrene [4]. Chinsakchai et al. graded severity of PCD as noncomplicated, impending venous gangrene and venous gangrene (Table 28.2) [4]. Patients with impending gangrene had changes in motor/sensory function with diminished pulses and skin blistering. Impending gangrene patients had a 22 % mortality and 7 % amputation rate compared to 0 % mortality and 20 % amputation rate in patients with uncomplicated PCD. Venous gangrene patients in their analysis had a 57 % mortality and 43 % amputation rate. The authors concluded that successful treatment outcome correlated with the initial graded severity of PCD and the surgeon’s experience, not the type of therapy.

The massive sequestration of fluid in the extremity with PCD can lead to shock as well as to the development of acute compartment syndrome. Patients present with mild to severe shock states with signs of pronounced volume depletion including hypotension, tachycardia, and low urine output that require aggressive fluid resuscitation. Investigators in the 1960s estimated that up to 6–10 L of interstitial fluid loss occurs within 5–10 days of onset of PCD [16]. Fluid sequestration in calf and thigh muscular compartments leads to elevation in resting compartment pressures, with pressures in PCD patients recorded as high as 70 mmHg [12, 13]. Elevated compartment pressures compound the ischemia present secondary to venous obstruction and are the rational for the selective use of fasciotomy in managing patients with PCD [12, 13, 15].

Venous gangrene develops in 40–60 % of patients with PCD with the time of onset usually between 2 and 8 days after developing PCD [3–6]. Gangrenous changes start distally with the majority of patients’ gangrene limited to the toes and foot (Fig. 28.2). Gangrenous changes can occur with intact Doppler signals due to thrombosis at the microcirculation level. This has been termed superficial gangrene and occurs in 10–20 % of cases [8]. Gangrene can extend more proximal to the foot involving the musculature of the calf and thigh. The development of venous gangrene has a grim prognosis. Interventions to treat patients presenting with venous gangrene cannot reverse the gangrenous process so they are designed to reverse the ischemia in the extremity and resuscitate the patient. Despite the growing use of less invasive catheter-driven techniques, outcomes of patients with venous gangrene have not improved as evidenced by the recent systematic analysis illustrating 57 % mortality and 43 % amputation rates [4].

Patients with PCD have a higher incidence of pulmonary emboli (12–40 %) than patients with submassive deep venous thrombosis [3, 6, 8]. Patients with venous gangrene have a higher incidence of fatal pulmonary embolism (21 %) than patients with PCD (3.4 %) [3, 22]. Pulmonary embolism is fatal in 50 % of patients with ischemia venous thrombosis and responsible for 30 % of deaths reported from PCD [4, 6]. A recent literature review found that thrombus in patients with PCD localized to the IVC in 35 % of patients and the iliac veins in 50 % [4]. It makes empiric sense that patients with large proximal thrombus burdens would be more likely to develop pulmonary embolism.

Post-thrombotic syndrome (PTS), characterized by the development of pain, edema, venous ectasia, and skin induration of the affected limb, develops within 2 years after an extremity DVT in 23–60 % of patients who have experienced extremity deep vein thrombosis [34]. Some authors have reported that proximal deep vein thrombosis and persistent venous obstruction are risk factors for the development of PTS [35, 36]. Information on developing PTS in patients with PCD or venous gangrene is limited; however, the incidence of PTS has been reported to range between 46 and 94 % [6, 31].

The highest incidence of ischemic venous thrombosis occurs in the fifth and sixth decades, but PCD has been reported in patients from age 6 months to 87 years [3, 37]. Early reports of series of patients with PCD found a higher incidence of PCD in women than men (4:3 ratio) [5]. More recent series of patients reflect an older demographic and reveal a move toward an incidence ratio of PCD in men versus women of 1.5–1 [4, 38]. The left lower extremity is affected more often than the right in patients with PCD [3, 38]. This is felt due to the development of elevated venous pressure secondary to compression of the left common iliac vein as it passes posterior to the right common iliac artery. May and Thurner first described the anatomic variant that bares their name in 1957 [39]. Approximately 22–24 % of the population is felt to have this compressive physiology [15]. The majority of patients with iliac vein compression however are asymptomatic but can have elevated venous pressures placing them at an increased risk for subsequent development of DVT.

Upper extremity ischemic venous thrombosis is rare compared to lower extremity accounting for 2–5 % of all reported cases of PCD/venous gangrene [2, 3]. Haimovici’s series of patients presenting with venous gangrene reported a 19 % incidence of upper extremity venous gangrene [3]. Clinical presentation mirrors the signs and symptoms of lower extremity ischemic venous thrombosis (Fig. 28.3). Patients with UEDVT who progress to ischemic venous thrombosis are more likely to have an underlying malignancy, hypercoagulable state, or low cardiac output condition [14, 40, 41]. Smith et al.’s retrospective review of the literature reported that the development of upper extremity venous gangrene was associated in 16 patients with either malignancy (31 %), hematologic abnormality (38 %), HIT (19 %), or IV infusions (44 %) and that 62 % of patients suffered from a severe systemic illness. Patients present with pain and cyanosis that begins in the fingers and progresses proximally. As PCD progresses to venous gangrene, arterial Doppler signals are lost and neurologic deficits ensue [2, 14, 40]. The majority of patients who present with venous gangrene have the combination of central and peripheral venous occlusion [40]. Patients who have central venous occlusion and low cardiac output states account for 12 % of patients. Unlike lower extremity PCD, a small number of patients with upper extremity PCD/venous gangrene do not have central vein (6 %) thrombosis but rather have a history of IV infusion in the affected extremity, peripheral vein thrombosis, and systemic illness causing a low cardiac output state [40]. The most central extent of gangrene has been reported to be the arm in 12 %, the forearm in 43 %, the hand in 19 %, and the fingers in 25 % [40]. Pulmonary embolisms have been reported in 19 % of upper extremity venous gangrene patients compared to only 1 % of patients with uncomplicated UEDVT [40].

The diagnosis of PCD and venous gangrene can usually be made based on the clinical findings of toxic-appearing patients with massive extremity edema, cyanotic to purple skin discoloration, and extreme pain in the extremity. However, further testing is necessary to confirm the diagnosis as well as to evaluate the extent of thrombus burden, assess for arterial insufficiency, and rule out compartment syndrome. A recent review of imaging modalities in a series of patients with ischemic venous thrombosis found that 37 % of patients underwent duplex ultrasound imaging (DUS), 34 % contrast venography, 15 % computed tomography angiography (CTA) , 2 % magnetic resonance venography (MRV), and 26 % a combination of studies [4].

Duplex ultrasound imaging is a rapid, inexpensive, reliable noninvasive way to confirm your clinical diagnosis. White et al. compared DUS against contrast venography for the diagnosis of proximal deep vein thrombosis and reported a sensitivity of 93 % and a specificity of 98 % for DUS [42]. Disadvantages of DUS include its operator dependence, decreased accuracy in obese patients, and decreased specificity in visualizing the extent of thrombus involving the IVC and iliac veins. Continuous-wave Doppler ultrasound is used for the calculation of ankle/brachial indices to quantify the presence of arterial insufficiency, while DUS color arterial imaging can be used in the PCD patient to noninvasively ascertain arterial anatomy and patency.

Contrast venography historically has been viewed as the “gold standard” to assess venous patency, but today it has been largely replaced by less invasive imaging. In cases of PCD, venography can be nondiagnostic in up to 20–25 % of cases [22]. Venography is expensive and can precipitate venous thrombosis in 2–3 % of cases [22]. Today, venous contrast venography is reserved for use in patients undergoing immediate catheter-directed thrombolysis (CDT), venous thrombectomy, or IVC filter insertion [4]. Contrast venography and CTA both utilize potentially nephrotoxic iodinated contrast whose potential harm is amplified by these patients’ severe volume-depleted state.

CTA and MRV are important adjunctive techniques that allow for the characterization of thrombus extension into the IVC, iliac, subclavian, and/or axillary veins while also potentially identifying the precipitating cause for venous thrombosis such as May–Thurner syndrome, Paget–Schroetter syndrome , pelvic mass, or other malignancy. CTA and magnetic resonance imaging have the advantage of also evaluating the arterial vasculature at the same time the venous anatomy is examined (Fig. 28.4).

Recognizing the development of acute compartment syndrome in patients with PCD is critical to their successful management. Leg compartment pressure measurements can be obtained using the wick catheter technique or more simply with the Stryker Intra-Compartmental Pressure Monitor System (Stryker Instruments, Kalamazoo, Michigan). This is a bedside procedure that should be performed early in patient management. Markedly elevated compartment pressures in patients with PCD have been reported to average 50 mmHg in the anterior and 48 mmHg in the deep posterior compartments of the affected leg compared to 6 mmHg in the anterior and 3 mmHg in the deep posterior compartments of the patient’s unaffected limb [13].