Chapter 104. Rheumatology in the ICU

• Most ICU admissions for rheumatology patients are prompted by infection.
• New-onset rheumatic diseases rarely prompt ICU admission in the absence of a revealing prodrome.
• In most patients without a previously established collagen vascular disease, suspected vasculitis will be explained by an alternative diagnosis.
• Immunoserologic assessment of critically ill patients is a double-edged sword providing both enlightenment and misleading shadows. All immunoserologic testing must be interpreted with a thorough understanding of the patient's clinical condition.
• Inability to assign specific diagnostic labels to patients with severe life-threatening immunoinflammatory disease should not delay therapeutic intervention.
• Not all ischemic skin lesions that appear to be vasculitis are. Pseudovasculitis of various causes should always be part of the differential diagnosis.
• Empiric trials with corticosteroids can be a rational approach to patient care when such trials are carried out appropriately.
• Acute organic brain syndrome without focal neurologic deficits or evidence of systemic vasculitis is unlikely to be secondary to vasculitis.
• Fever in patients with collagen vascular disease should be presumed infectious if accompanied by chills, leukocytosis with a left shift, or hypotension.
• Patients who have been treated with significant doses of corticosteroids within the past year may require empiric replacement therapy during critical illness or surgical procedures until adrenal insufficiency can be excluded.

Rheumatologists seldom lend their expertise to the medical ICU. The tempo and level of intensity of most rheumatic disorders are more suited to outpatient management. Nonetheless, approximately 10% to 25% of all rheumatic disease patients visiting the emergency department require hospitalization and up to one-third of these patients will require intensive care.1 Patients with rheumatic diseases admitted to the medical ICU most often have problems not directly related to their primary illness. Sepsis, massive gastrointestinal bleeding, and myocardial infarcts may arise secondary to treatment. The major direction of care in these circumstances often comes from the intensivist. Circumstances do arise that require the unique insight of the experienced clinical rheumatologist, who at times must direct the management of a disease-specific complication. Just as often the rheumatologist is asked to address a diagnostic dilemma spawned by puzzling clinical and laboratory data. This chapter addresses the more common issues that prompt the rheumatologist and the intensivist to collaborate.

Systemic Lupus Erythematosus

Fever: Is It the Lupus?

Fever in the patient with lupus presses the clinician for an urgent answer to the question: Is this caused by lupus activity or infection? Fever is a common finding in active systemic lupus erythematosus (SLE) and follows no specific curve or pattern.2 It may respond to the usual antipyretics or require corticosteroids. Single-daily-morning dose prednisone may not control late afternoon or evening fevers. Leukocytosis and increased bands on peripheral smear are strong presumptive evidence for infection, as is the presence of shaking chills. Complement proteins or components, including C3 and C4, are acute-phase reactants and usually rise with infection. Low levels of complement occur in some but not all patients with active lupus. Using discriminant analysis, Inone and colleagues3 showed that 95% of 74 febrile episodes could be correctly classified as to the cause of fever when a combination of white blood count (low in SLE, normal to high with infection) and α2-globulin levels (high with SLE, normal with infection) are used as variables.

In the ICU setting, the febrile patient with SLE is probably best considered infected and treated with broad-spectrum antibiotics pending results of cultures.4 Infection is most likely to be caused by nonopportunistic organisms, and coverage for gram-positive and gram-negative aerobes represents adequate empirical therapy when no obvious source has been recognized. Systemic infections with Salmonella, endocarditis involving lupus-related valvular lesions, and pneumococcal sepsis in the splenectomized (surgical or autosplenectomy) are among the infections that have special significance for lupus patients.

Myocardial Infarct in the Lupus Patient: Is It Vasculitis?

Myocardial infarction makes a major contribution to excessive mortality in SLE. Mortality data suggest a bimodal distribution with many late deaths in SLE secondary to ischemic heart disease. The primary cause of ischemic heart disease in lupus is atherosclerosis. Autopsy studies in young women with lupus with death from any cause show an excessive incidence of atherosclerosis in comparison to age-matched females.5 The cause of accelerated atheroma is not totally clear, but inflammation of the vascular endothelium due to chronic immune complex disease compounded by the effects of corticosteroids probably is the major contributing factor.

Vasculitis in SLE may affect any organ including the heart. Most patients with coronary artery vasculitis have evidence of vasculitis in other organs. Serologic evidence of active lupus and markers of systemic inflammation including low hemoglobin and albumin are likely to be present. Acute therapy would include high-dose corticosteroids (see below). In the face of widespread vasculitis, additional intervention with parenteral immunosuppressives, B-cell–depleting biopharmaceuticals, and plasmapheresis may be indicated.

Another potential cause of myocardial ischemia is coronary thrombosis secondary to antiphospholipid syndrome.6 Treatment is anticoagulation with heparin. Thrombolytic therapy has been successfully used in some patients with antiphospholipid syndrome.7

Corticosteroids are potentially hazardous, and high doses may predispose to ventricular rupture. Therefore corticosteroids should not be used in the absence of reasonable clinical suspicion for coronary vasculitis. If the lupus patient is in apparent clinical remission without evidence of widespread vasculitis, treatment should be the same as that for the non-lupus patient. Early cardiac catheterization may help clarify the nature of the vascular disease.

Renal Failure: Is It Treatable Lupus Nephritis?

In patients with SLE in the ICU, renal insufficiency may be caused by a variety of factors, including drugs, especially nonsteroidal anti-inflammatory drugs (NSAIDs), hypovolemia, sepsis, or previous renal disease.8 In some cases, active lupus nephritis is a contributing factor. A careful examination of the urinary sediment is the most critical diagnostic tool. Proteinuria, casts, and red blood cells indicate glomerulitis. Lupus patients with active nephritis are most often hypertensive. Significant renal lupus (other than membranous disease) is often associated with low complement levels and elevation of anti-DNA antibody.

In a patient with a creatinine level above 4.0 mg/dL who has been adequately hydrated, has been divorced from nephrotoxic drugs, and shows evidence of active glomerulitis, the question arises: Is more aggressive immunosuppression desirable? The answer to this question depends on the degree of potential disease reversibility. A renal biopsy can help clarify this issue. The presence of significant chronic disease should dampen enthusiasm for aggressive therapy. Review of old records can be enlightening if long-standing loss of renal function is documented.

Clinicians have become increasingly aware that immunosuppression in the lupus patient with advanced renal disease can be more hazardous than progression to complete renal failure. Patients with lupus tolerate dialysis in a fashion comparable to other patients, and results of renal transplantation are favorable.9 Paradoxically, patients with lupus who develop chronic renal failure often enjoy an amelioration of extrarenal symptoms.10 A few patients have recovered sufficient renal function to allow withdrawal from dialysis. For all these reasons, the overzealous administration of immunosuppression in patients with lupus and advanced renal disease is to be discouraged.

Respiratory Failure and Lung Infiltrates: Is It Lupus Pneumonitis?

Respiratory failure in the patient with SLE is an ominous development. A paradigm of a compromised host is a patient with lupus who is on high-dose corticosteroid therapy. The usual opportunistic pulmonary infections need to be urgently excluded by bronchoalveolar lavage, bronchoscopic transbronchial biopsy, or open lung biopsy. If no superimposed infections or embolic etiology can be found and treated, lupus-related respiratory failure remains a diagnosis of exclusion and can be caused by either lupus pneumonitis or diffuse pulmonary hemorrhage. Acute lupus pneumonitis may occur as an initial manifestation of SLE and is characterized by fever, tachypnea, and hypoxemia, which may be accompanied by cough, pleuritic chest pain, and hemoptysis.11 Radiologic findings are highly variable but usually bilateral and at least bibasilar. This diagnosis is not only one of exclusion; it unfortunately still rests solely on clinical suspicion.

The other SLE-related cause of respiratory failure is diffuse alveolar pulmonary hemorrhage.12 Although invasive Aspergillus or tuberculosis can erode a pulmonary vessel and cause hemorrhage, gross hemoptysis, when present, usually indicates alveolar hemorrhage. Hemoptysis is not usually seen in lupus pneumonitis; unfortunately, this finding is present in <50% of patients with alveolar hemorrhage. Blood or hemosiderin-laden macrophages found during bronchoscopy in a patient without heart failure can be helpful findings but are nonspecific. The presence of thrombocytopenia is not helpful, but bleeding sufficient to cause acute respiratory failure most invariably causes an acute drop in hematocrit. In fact, treatment should not be delayed in order to distinguish between lupus pneumonitis and lupus-associated hemorrhage, because mortality is extremely high in either syndrome and treatment strategies are similar.

Pulmonary hypertension, sometimes severe, is frequently present. Cardiac filling pressures, in contrast to B-type natriuretic peptide determinations, can occasionally be helpful discriminators to exclude acute cardiogenic pulmonary edema. Pulmonary artery thrombosis masquerading as massive pulmonary emboli can occur in patients with pulmonary hypertension or thrombotically active anticardiolipin (ACL) antibodies.

The mortality of lupus pneumonitis is high, and treatment should be aggressive. Individual preferences will dictate modes of therapy, since no consensus exists on either the etiology of the syndrome or effective treatment. Pulse methylprednisolone, usually 500 to 1000 mg intravenously given for 3 to 5 days, or bolus cyclophosphamide at 0.5 to 1.0 g/m2 has been used. Uncontrolled studies of plasmapheresis for alveolar hemorrhage in SLE do not reveal striking therapeutic efficacy. In the absence of alternate therapy with proven effect, plasmapheresis may be useful as a temporizing measure during the induction period required for the optimal effects of immunosuppressive agents like cyclophosphamide.13

Brain Dysfunction: Is It Lupus?

A patient with acute, severe neurologic deficits and a history of SLE or a clinical syndrome and laboratory evidence suggestive of systemic vasculitis presents a diagnostic and therapeutic dilemma for the critical care clinician. In the vast majority of instances, involvement of the central nervous system (CNS) in SLE occurs as an organic brain syndrome or seizures in young females with active multisystem disease. Patients with organic brain syndromes manifest psychosis usually early after the onset of SLE. A common dilemma is to differentiate between steroid-induced pseudo-organic brain syndromes and those owing to active SLE. Interleukin-6 (IL-6) levels are markedly increased in the CSF of patients with active CNS SLE. Minimal, if any, IL-6 should be present in the CSF of patients with steroid-induced psychosis, and IL-6 may prove very helpful in differentiating between drug-induced CNS changes and active SLE. Early-onset development of psychosis while on low doses of corticosteroids and fulminant progression usually characterize the CNS disease caused by SLE.14 Depressive syndromes with or without auditory, visual, or even olfactory hallucinations in a patient on a dose of corticosteroids usually higher than 0.5 mg/kg per day are usually drug-induced. Because there is little evidence that high-dose corticosteroids ameliorate the CNS manifestations of SLE, rapid reduction of corticosteroid dosage is usually the simplest and most direct clinical strategy when a quandary exists as to whether CNS dysfunction is drug-induced.

Lupus-induced seizures are usually grand mal but can be very pleomorphic and usually occur in patients with active SLE.15 Rapidly progressive, multisystem SLE that is further complicated by early onset of a seizure syndrome usually portends a poor outcome. The development of a seizure disorder in a patient with a long history of inactive SLE or in a patient with SLE entering remission is usually caused by something else. There is also little clinical evidence that high-dose corticosteroids alone influence the course of seizure disorders in SLE, and these patients should be aggressively treated with conventional anticonvulsive therapy. A difficult but critical differentiation must be made in SLE patients with CNS findings, anemia, and thrombocytopenia. The latter findings, while very common in active SLE, coupled with peripheral blood smear evidence of microangiopathy, point instead to thrombotic thrombocytopenic purpura (TTP) (see Chap. 70). Case reports suggest that TTP and SLE can co-occur, and the differentiation between the two diseases is vital since the treatment of life-threatening TTP is plasma exchange and not concomitant pulse methlyprednisone, alkylating agents, and/or plasmapheresis that many rheumatologists, in spite of unproven benefit, will resort to in the setting of fulminant CNS SLE.

Three other CNS manifestations of SLE can be puzzling. A small subset of patients with SLE who have taken NSAIDs, especially ibuprofen, will develop a meningitis-like picture that is characterized by fever, severe headache, nuchal rigidity, and cerebrospinal fluid neutrophilia or pleocytosis.16 In an immunosuppressed patient, these findings prompt consideration of both common and unusual bacterial and fungal etiologies. The entire syndrome will remit rapidly once the drug is discontinued.

Migraine headaches, some of spectacular intensity, are frequent in SLE.17 They usually respond to increases in corticosteroid dosage.

Actual paralysis is uncommon in SLE. Rare instances of transverse myelitis occur in the context of active SLE, and myelopathy may rarely be the presenting symptom of lupus. Until recently this was thought to be secondary to necrotizing vasculitis in the spinal arteries. It is likely, however, that many instances have been secondary to thrombosis associated with high titers of IgG ACL antibodies.18 This differentiation may be important because the primary treatment of this syndrome is not aggressive immunomodulation of SLE, but appropriate anticoagulation.

Some investigators have reported elevated serum levels of anti-ribosomal P antibodies in patients with lupus-induced depression or psychosis. Other studies have not confirmed the association of these antibodies with psychosis.

Scleroderma

Pulmonary Hypertension

The emergence of effective therapeutic options has given the detection of pulmonary vascular disease a new sense of urgency. The exact prevalence of pulmonary hypertension in scleroderma is unknown. Hospital-based studies would suggest the number to be in the 10% to 15% range.19 There are generally two settings in which it is identified. In patients with limited cutaneous systemic sclerosis (calcinosis, Raynaud's phenomenon, esophageal motility disorders, sclerodactyly, and telangiectasia [CREST] variant of scleroderma) it occurs classically as an isolated phenomenon in the absence of pulmonary fibrosis. This generally occurs in the second decade of disease or later. Patients who fall into the second major category of scleroderma, diffuse disease, may develop pulmonary hypertension as the result of advanced pulmonary fibrosis. In both settings the vascular disease is characterized by bland endothelial proliferation and vascular occlusion. The vasculopathy of scleroderma is not characterized by an inflammatory infiltrate and is not treated with corticosteroids or immunosuppression in the manner that typifies the treatment of lupus vasculitis.

Early symptoms of pulmonary hypertension are exertional breathlessness, but later symptoms are likely to provoke admission to the medical ICU or critical care unit (CCU). These include near-syncope or syncope with exertion and right ventricular angina or heart failure. Scleroderma patients, especially those at greatest risk for pulmonary hypertension, are now screened yearly or every few years using Doppler echocardiography. Confirmation of diagnosis demands a right heart catheterization, which provides additional important information about pulmonary capillary wedge pressure and cardiac output.

The treatment of pulmonary hypertension in the context of scleroderma is similar to that in idiopathic and familial pulmonary hypertension.20 Anticoagulation, diuretics, digoxin, and supplemental oxygen form the basic management program. A small subset of patients may benefit from calcium channel blockers. Two exciting breakthroughs have substantially changed the quality of life of these patients for the better. Parenteral epoprostenol was initially viewed as a bridge to transplant, but some patients have elected to continue the drug as a chronic maintenance therapy. The emergence of a broad-spectrum endothelin receptor antagonist in the form of bosentan now offers the option of oral therapy to patients with moderately severe disease.20 End-stage pulmonary hypertension may require heart-lung transplantation. Traditionally scleroderma patients have been considered poor candidates for transplantation. However, a recent report of 12 patients with scleroderma-associated pulmonary artery hypertension who underwent transplantation suggests similar beneficial outcomes compared to those with other acquired lung diseases.21

Other less common causes of pulmonary hypertension should not be overlooked in scleroderma patients, including recurrent thromboembolic disease. This may be a particular problem in patients with antiphospholipid antibodies.

Hypertensive Renal Crisis

Until recently, hypertensive renal crisis in patients with scleroderma was the major cause of mortality. The advent of angiotensin-converting enzyme (ACE) inhibitors thrust pulmonary disease into the role of leading cause of death in scleroderma. Nonetheless even recent studies of patients with scleroderma renal crisis suggest a continued high mortality rate in this subset of patients.22 Therefore strategies for prevention of scleroderma renal crisis are crucial. Scleroderma renal crisis typically develops in patients with diffuse cutaneous disease and rarely in patients with limited cutaneous disease or CREST.23 At its onset patients experience a marked increase in blood pressure that may be accompanied by abnormalities of urinary sediment (erythrocytes and protein) and the peripheral blood smear (fragmented cells) and thrombocytopenia. Headache, visual disturbance, congestive heart failure, and mental status dysfunction may accompany the hypertension.

The pathogenesis of hypertensive renal crisis is complex and involves a very high renin state. Although combinations of older antihypertensive agents were occasionally successful, the advent of ACE inhibitors has revolutionized the outlook for this problem. Rheumatologists have a low threshold for using these agents in scleroderma patients and typically initiate them at the first diagnosis of hypertension.

In the setting of acute hypertensive renal crisis, larger doses of ACE inhibitors should be used. Although captopril, enalapril, and newer ACE inhibitors have all been used, some prefer captopril because the dose can be adjusted more flexibly. Angiotensin receptor blockers, calcium channel blockers, prostacyclin, and endothelin receptor antagonists have all been suggested as adjunctive therapy for patients refractory to maximum doses of ACE inhibitors.24

Some patients may progress to complete renal failure despite blood pressure control. Continued use of ACE inhibitors and dialysis may be required for months prior to recovery of renal function. Such improvement can continue for up to 2 years.

Progressive renal failure can occur in the absence of significant hypertension in as many as 10% of patients with scleroderma renal crisis. Microangiopathic changes may be seen on peripheral smear. Treatment with ACE inhibitors is indicated for this normotensive subset of patients.

Polymyositis/Dermatomyositis

Diagnosis in the ICU

Very ill patients in the ICU may be weak and have elevations of creatine phosphokinase (CPK). Such clinical data prompt speculation about the presence of immune-mediated myositis. The most common presentation of polymyositis is the insidious onset of proximal muscle weakness, at times associated with myalgia. The acute development of de novo polymyositis in the ICU is unlikely. Similarly, acute fulminant disease requiring ICU admission with subsequent diagnosis is uncommon. Nonetheless, patients with undiagnosed polymyositis may be discovered in the ICU following admission for another reason (e.g., aspiration pneumonia). More likely is the presence of weakness (usually generalized) in combination with a spurious or nonimmune cause of CPK elevation. Intramuscular injections and myonecrosis during severe episodes of hypotension are common causes of increased CPK levels in critically ill patients. ICU-acquired myopathy can usually be distinguished from polymyositis by clinical history. Critical illness myopathy is usually characterized by normal or modestly elevated serum CPK levels.25

The skin lesions of dermatomyositis are so highly characteristic as to be diagnostic of dermatomyositis when accompanied by weakness and an elevated CPK value. Nearly all patients with active polymyositis will display an elevated CPK or aldolase level, although occasional patients will have normal muscle enzymes.26 A unilateral electromyogram (EMG) can provide supportive evidence for the presence of myopathy and identify a biopsy site. Fibrillation potentials suggest active inflammation. A bedside EMG can be done in the ICU, although technical artifact may complicate the interpretation. Magnetic resonance imaging (MRI) may also confirm the presence of inflammatory muscle disease, but is generally impractical for ICU patients.

The EMG should be done unilaterally because EMG needle artifact may be confused with muscle inflammation histologically. Because polymyositis/dermatomyositis is a symmetrical disease, the corresponding maximally affected muscle group can be biopsied on the opposite side. Open biopsy can be done at the bedside, preferably by an experienced surgeon to ensure proper handling of muscle tissue. Needle biopsy of muscle using a Polley-Bickel needle is advocated by some and appears suitable for providing confirmation of active muscle inflammation.

Respiratory Failure

Patients with polymyositis/dermatomyositis may develop respiratory failure secondary to muscle weakness involving the diaphragm, intercostals, and accessory muscles. If pharyngeal muscles are involved, acute respiratory failure may be precipitated by aspiration pneumonia. Patients with respiratory failure have a poor prognosis.27 Some patients with dermatomyositis and this type of profound weakness harbor an underlying malignancy. Such patients are notably refractory to treatment.

Steroids are the mainstay of acute management of inflammatory myositis. Prednisone, 1 to 2 mg/kg per day or its approximate intravenous equivalent of methylprednisolone (in single or divided doses) may be given. In the ventilator-dependent patient, a short trial of pulse steroids may be justified—500 to 1000 mg methylprednisolone intravenously every day for 3 days. Improvement in respiratory muscle strength can be judged by a rise in the maximal inspiratory pressure. Intravenous immunoglobulin (IVIg) is another option for the acute management of the severely ill patient refractory to therapy with corticosteroids.28 Second-line agents in polymyositis/dermatomyositis include methotrexate, azathioprine, and cyclophosphamide. Azathioprine is not useful in the acute setting owing to slow onset of action. Methotrexate is most commonly used after corticosteroids. In the ICU setting the intravenous route of administration may be preferred; the dose is 25 to 75 mg intravenously weekly.

Uncontrolled trials with plasmapheresis and combination plasmapheresis and leukocytapheresis have reported benefits and may be considered in refractory patients.29,30 However, a randomized double-blind trial comparing plasmapheresis, leukapheresis, and sham pheresis in 39 patients with polymyositis and dermatomyositis showed comparable improvements in the three treatment groups.31 The role for therapeutic pheresis in patients refractory to conventional therapy seems limited.

Rheumatoid Arthritis

Methotrexate Pneumonitis

Oral low-dose methotrexate given intermittently emerged as the major therapeutic innovation in the treatment of rheumatoid arthritis (RA) during the 1980s. In the era of biological therapies methotrexate continues to be the gold standard of RA therapy and a frequent adjunct to cytokine-directed therapy. Methotrexate therapy is both highly effective and generally well tolerated. A major, albeit uncommon, toxicity is an acute pneumonitis characterized by dyspnea and nonproductive cough.32,33 Fever is a frequent accompaniment.34 Diffuse alveolar and interstitial infiltrates are present at diagnosis or appear within days. Opportunistic infections mimicking this syndrome have only rarely been reported.35 According to one study risk factors for the development of methotrexate-induced lung injury include elderly patients, diabetes, and underlying rheumatoid pleuropulmonary disease.36 Patients suffer profound hypoxemia. They appear extremely ill, and deaths have been reported. The mechanism is unclear but is presumed to be a hypersensitivity reaction to the drug. Some patients have been rechallenged without developing the syndrome while others have relapsed.

Diagnosis depends on the above clinical scenario developing in a patient taking methotrexate at any dose. Duration of treatment prior to symptoms has been variable. Bronchoscopy with brushings and biopsy shows nonspecific inflammation, and bronchoscopy's main justification is to rule out an opportunistic infection. Because these are rare, it is not unreasonable to forego bronchoscopy initially. Open-lung biopsy is usually unnecessary.

Treatment includes O2, withdrawal of drug, and corticosteroids. Some have argued that steroids are not critical to recovery. The usual dose is prednisone 1 mg/kg per day or its equivalent in single or divided doses. Most patients will show signs of recovery within a week. Eventually, recovery is typically near complete.

Cervical Spine Subluxation

Arthritis commonly affects the cervical spine in rheumatoid arthritis, with estimates as high as 80% of patients. Subluxation of vertebrae secondary to ligamentous laxity may occur at single or multiple levels. Anterior atlantoaxial subluxation of C1 on C2 is the most frequent cervical abnormality and is particularly dangerous because of the capacity of the odontoid process (or dens) of C2 to compress the anterior spinal cord with motion. Thus sudden hyperextension of the neck during intubation could result in quadriplegia. In reality such occurrences are rare. The explanation may in part include the fact that progressive resorption of the dens often accompanies the most severely unstable necks. Symptomatic patients can be diagnosed with MRI or a myelogram. However, some dramatic subluxations on MRI are not accompanied by neurologic signs or symptoms. Flexion and extension films of the cervical spine may show dynamic instability and subluxation of C1 on C2. There are few data about the specificity or sensitivity of such films to predict a cervical cord catastrophe. Clearly, caution should be exercised in the intubation of patients with rheumatoid arthritis and neck disease; if time allows, nasotracheal or fiberoptically guided endotracheal intubation is preferred. However, when the risk of delayed intubation is sufficiently great, the procedure should not be delayed for radiographic studies, because the risk of cord catastrophe is quite low. Cervical instability is a problem that attends advanced destructive rheumatoid arthritis. Early aggressive treatment of RA patients with disease-modifying therapy has dramatically reduced the prevalence of this problem.37

Outcomes of Patients with Rheumatologic Diseases in the Intensive Care Unit

Mortality rates for patients with systemic rheumatic diseases admitted to the ICU is high.38 This is perhaps not unexpected when consideration is given to a typical scenario of a chronically ill patient with multiple impaired organs, a disordered immune system, treatment with chronic immunosuppression, and admission to the ICU with infection. Reported ICU mortality rates for patients with rheumatic disease are higher than would be predicted using the Acute Physiology, Age, and Chronic Health Evaluation II or simplified acute physiology score II (SAPS).39,40 Factors associated with poor outcome include higher SAPS II scores, poor health status prior to admission, duration of rheumatic disease, corticosteroids and immunosuppressive drugs, renal failure, coma, and acute respiratory distress syndrome. The overall ICU mortality rate for patients with rheumatic disease ranges from 30% to 60%.38 Mortality rates are higher for patients admitted because of infection compared to those admitted for exacerbation of rheumatic disease.39,40 Since infection may cause two-thirds of ICU admissions among patients with rheumatic disease, aggressive intense therapy of the febrile patient with broad-spectrum antibiotic coverage is appropriate.

Fever of Undetermined Origin: Rheumatic Causes

The traditional definition of a fever of undetermined origin (FUO) depicts a patient with a significant fever of 6 weeks' or greater duration and no definable cause. In actual practice, depending on the impatience of the attending physician, FUO usually becomes the working diagnosis within 3 to 14 days after a fruitless search for classic causes of pyrexia. The most common causes by far of FUO are occult infection, drugs, and occasionally malignancy. After those are excluded, one must consider a limited array of rheumatic diseases that could be present in a febrile patient with protracted fever. Rheumatic disorders may account for 10% of cases of FUO. Because rheumatoid arthritis, scleroderma and its variants, dermatomyositis, polymyositis, and polymyalgia rheumatica are not usual causes of significant fever, they need not be strongly considered.

SLE can present with high fevers, either spiking or relatively constant, and leukopenia, hypoalbuminemia, anemia, and an elevated erythrocyte sedimentation rate (ESR), but few other overt clinical signs of lupus such as rash or serositis, polyarthritis, or active urinary sediment. Since SLE is readily diagnosed serologically, it represents a small fraction of FUO patients. Most systemic necrotizing vasculitides will be evident after examination of the skin, chest radiograph, and urinary sediment. Myocardial dysfunction, age, and demographics of the patient with fever may point to either Kawasaki's disease or Lyme disease as the source, and acute valvular dysfunction is being increasingly recognized in SLE and may be the clue to acute rheumatic fever in an adult.41 An extremely high ESR is a nonspecific laboratory clue in FUO, but antibodies to relevant specific antinuclear or streptococcal disorders provide supportive evidence.

A most vexing diagnosis to pin down is that of adult Still's disease. These patients will have relentless spiking fevers, at times a history of FUO in childhood, leukocytosis, mild to moderate hepatic enzyme changes, and an occasional truncal rash.42 Polyarthritis or arthralgias are not a constant feature early in this syndrome. Hepatosplenomegaly and lymphadenopathy are common. Rarely, acute pericarditis with tamponade or myocarditis with respiratory failure may complicate the course of adult Still's disease.

In the final analysis, treatment may have to be based on the supposition that the patient has relentless, immunologically driven, noninfectious inflammation that evades specific diagnosis. Blanket suppression of cytokines and white blood cell responses by corticosteroids may be necessary. The steroid dose is tailored to control fever and normalize the acute-phase response. NSAIDs may provide a less potent but possibly better tolerated alternative strategy.

Multiple Autoantibodies and Multisystem Inflammatory Disease: What Name Do I Give It?

The collagen-vascular diseases are characterized by the presence of sterile inflammation in multiple organs and multiple autoantibodies.43 The prototype disease is SLE, which is characterized by the widest clinical and serologic spectrum. Other diseases include scleroderma, Sjögren's syndrome, polymyositis dermatomyositis, rheumatoid arthritis, and syndromes with overlapping features (overlap syndrome or undifferentiated connective tissue disease). One subset of the latter group has been called mixed connective tissue disease (MCTD). The clustering of clinical features and the nature and diversity of autoantibodies may strongly suggest one disorder rather than another, but the overlap of clinical, serologic, and pathologic features among these diseases is great. This can lead to considerable consternation for clinicians. Debates about whether a given patient has SLE, primary Sjögren's syndrome, MCTD, or another overlap syndrome are tiring, usually unresolvable, and generally irrelevant. The therapy for the immunologically active phase of these disorders is not disease-specific. The absence of a consensus label should not delay therapeutic efforts.

The Elderly Patient with an Elevated Sedimentation Rate: Is This Temporal Arteritis?

Within the ICU, advanced age, elevation of the ESR, anemia, and various nonspecific clinical features—chiefly fever—may converge to raise the question of temporal arteritis. Temporal arteritis is a granulomatous vasculitis that affects those over age 60 (mean age 70) and has a proclivity to affect extracranial vessels and branches of the aorta. Myriad uncommon clinical features may develop, but headache, polymyalgia rheumatica, visual disturbance, scalp tenderness, and jaw claudication are the common fingerprints of this disease. Treatment is highly effective with prednisone in doses of 40 to 60 mg/d.44

The onset of this disease is typically insidious, and its complications rarely prompt admission to the ICU. Therefore, the question is more commonly framed in terms of whether the patient has developed new-onset temporal arteritis in the ICU (probably not) or have it as a comorbid state at admission? History and old records are critical in this regard.

Physical examination is unreliable. Temporal artery biopsy is the gold standard, although sensitivity varies with the institution. A classical clinical scenario that includes headache, visual symptoms, jaw claudication, scalp tenderness, and polymyalgia rheumatica may be sufficiently compelling to prompt treatment. Usually the situation is murkier. There is no substitute for the temporal artery biopsy, which can be done under local anesthesia with little morbidity. A sufficiently large piece (4 to 5 cm) should be obtained and adequate cuts done for pathology. Contralateral biopsy is done routinely by some if a first biopsy is negative. That decision will be influenced by the details of the clinical scenario and the risk of an empiric trial with steroids. A negative biopsy in a marginal clinical situation is reasonable grounds to withhold therapy.

Abdominal Pain and Elevated Erythrocyte Sedimentation Rate: Is This Vasculitis?

The clinical presentation of acute abdominal pain is constrained to a limited number of clinical findings, regardless of cause. Unfortunately, there is nothing unique about massive gastrointestinal bleeding or the development of an acute abdomen when a patient has a high ESR and vasculitis is the etiology.

If the patient is already known to have systemic vasculitis, and this is usually the case when the gut becomes involved, the differential diagnosis of the acute abdomen and elevated ESR is a bit more simple. Here, early gastroduodenoscopy and colonoscopy with biopsy and perhaps vascular contrast studies may be necessary. MRI can assess arterial wall thickening, and in large-vessel vasculitis magnetic resonance angiography correlates well with conventional angiography. In the rare situation when a patient with no or relatively limited vasculitic findings in skin and renal and pulmonary systems develops fulminant vasculitis of the bowel, diagnosis may remain elusive until arteriography or surgical exploration. Patients on high-dose corticosteroid therapy for pre-existing systemic vasculitis may develop small-bowel or colonic ischemia or perforation with relatively few physical findings, even in the face of profound intra-abdominal sepsis.

Unusual causes of acute abdominal pain in patients with systemic vasculitis include hepatic infarctions, especially in SLE and polyarteritis, which can mimic many other intraperitoneal acute events. In this situation, hepatic computed tomography (CT) or MRI imaging is helpful. Massive bleeding can occur in most forms of systemic vasculitis, especially Henoch-Schönlein purpura, polyarteritis, and mixed cryoglobulinemia. Diarrhea and profound protein-losing enteropathy and acute bowel obstruction secondary to adhesive serositis have occurred rarely in SLE. There is nothing unique about acute pancreatitis in the patient with SLE; appropriate biochemical and imaging assessment will confirm this diagnosis.

The proper perspective must be maintained when the clinician attempts to make the diagnosis of systemic vasculitis involving the gut by analysis of small bowel or colonic tissue, since <50% of endoscopically obtained biopsies will be helpful in the diagnosis.45 Although careful circumspection should be exercised before embarking on exploratory laparotomy in patients with SLE or other vasculitis with abdominal pain and suspected perforation, one should err on the side of surgical exploration. More than one-half the patients with SLE and an acute abdomen will die.46 Surgical exploration will correctly identify the major pathophysiology of the syndrome and dictate appropriate management. Increasing availability of laparoscopes with enhanced optics may obviate the need for laparotomy. The vast majority of patients with polyarteritis nodosa who develop an acute abdomen usually die from extensive bowel ischemia, infarction, and perforation unless early surgical intervention is done. The contemporary use of alkylating agents, corticosteroids, and other aggressive therapies may reduce this highly fatal complication in this group of vasculitides.

In summary, any patient with vasculitis, especially one with SLE, who develops an acute abdomen is in dire jeopardy. Routine clinical, laboratory, and radiology assessments are not usually helpful. Early endoscopy, biopsy, arteriography in selected instances, and ultimately early surgical exploration are usually necessary.

Lung Infiltrates in Renal Failure: Is This an Immune-Mediated Pulmonary-Renal Syndrome?

When a patient, usually but not always a young male, has abnormalities of renal function and then develops increasing shortness of breath, hemoptysis, and bilateral diffuse infiltrates, an immune-mediated pulmonary-renal syndrome should be strongly considered and ruled out as soon as possible. The clinician must be wary because the symptoms can occur in reverse; that is, increasing compromise of pulmonary function may precede significant changes in renal function.

Hemoptysis has many causes in patients with renal failure. Some examples would include SLE, Wegener's granulomatosis, advanced cardiac failure, microscopic polyangiitis, infection, and rapidly progressive glomerulonephritis. The latter disease in actuality is probably one end of the spectrum of a group of diseases mediated by antibodies to epitopes common to both alveolar and glomerular basement membranes (GBM).47 In its classical expression, anti-GBM disease is called Goodpasture's syndrome.

Most diseases with pulmonary and renal manifestations can be distinguished from the immune-mediated pulmonary renal syndromes by their unique extrarenal characteristics. Once they are excluded, one is then confronted with a patient with glomerulonephritis and pulmonary hemorrhage. The diagnosis of the pulmonary-renal syndrome rests on establishing the presence of anti–basement membrane antibodies in the peripheral blood or in situ deposition in a renal biopsy. Characteristic histologic findings in the renal biopsy are those of a diffuse proliferative necrotizing glomerulitis highlighted by a somewhat unique characteristic of rather exuberant crescent formation. Although not well documented by contemporary clinical studies, immunohistologic analysis of tissue obtained by the transbronchial route has not been helpful in most instances.48 The diagnosis of an antibody-mediated pulmonary-renal syndrome is important because efficacious therapy exists, especially when initiated early in the disease course. Alveolar GBM antibodies can be rapidly removed by vigorous plasmapheresis, and then B-cell clones with basement membrane specificities can be suppressed or clonally eliminated with cytoreductive therapy.49 The efficiency of GBM antibody removal can be monitored easily by sequential GBM antibody assays. The antibody-mediated pathogenesis of Goodpasture's disease is beyond doubt and makes this syndrome a logical candidate for B-cell attrition with the chimeric monoclonal anti-CD20 antibody rituximab.50 The approach to a patient in the ICU with hemoptysis and renal failure should include close attention to maintenance of adequate pulmonary function, rapid acquisition of tissue (either renal or pulmonary or both), and appropriate immunopathologic analysis of any and all biopsy specimens in parallel with serologic assays for circulating GBM antibodies.

CNS Dysfunction: Is This Cerebral Vasculitis?

Patients with or without connective tissue disease and nonspecific markers of systemic inflammation who develop signs of brain dysfunction are routinely suspected of having CNS vasculitis. Often the likelihood is considered low but addressed in the interest of being thorough.

Rarely does CNS vasculitis cause psychosis or coma without focal neurologic signs. Primary angiitis of the CNS, particularly in early stages, is an exception, but would not be expected to show signs or symptoms of multisystem disease or coma that would prompt ICU admission. Finding focal deficits on examination is problematic in the comatose or disoriented patient. A skilled neurologist can contribute more than imaging techniques. MRI and CT scanning may reveal a lesion suggesting an ischemic event. While no pattern is specific for vasculitis, a completely normal MRI makes the diagnosis less likely. Angiography is the gold standard, although in small-vessel vasculitis angiography may be nondiagnostic. Access to angiography can be problematic for the critically ill ICU patient. Leptomeningeal biopsy is the gold standard for diagnosis of granulomatous angiitis of the CNS, an important subset of CNS vasculitis. An empiric trial with steroids may be appropriate.

Hypersensitivity Vasculitis: What Is the Cause?

The most common form of cutaneous vasculitis in any setting is leukocytoclastic vasculitis (LCV), often referred to as hypersensitivity vasculitis and recognized clinically as palpable purpura. The palpability of these lesions is caused by inflammatory infiltrates. Other less common expressions include bullous disease and cutaneous ulcers. The lesions have a tendency to occur in dependent areas and in ambulatory patients are found in the legs and buttock area. Localization may be atypical in bedridden ICU patients. By far the most common cause of new-onset LCV in the ICU patient is drugs. Any drug may be implicated, but thiazides, sulfa drugs, and penicillins are common offenders. Clues to nondrug causes would include significant digital involvement including associated nailbed infarcts, ulcerative lesions, significant glomerular inflammation, and any other evidence of extracutaneous vasculitis such as mononeuritis multiplex.

The presence of cryoglobulins or low complement levels suggests a nondrug cause. Skin biopsy can be done early but usually adds no additional diagnostic information. Predominant lymphocytic infiltration may help to support a diagnosis of drug-induced disease.51 Immunofluorescence on biopsied tissue is rarely helpful, but the presence of IgA is supportive of a diagnosis of Henoch-Schönlein purpura.52 Progression to major vasculitis does not predictably occur, but hypersensitivity vasculitis may be the presenting manifestation of a major systemic vasculitis such as Wegener's granulomatosis or Churg-Strauss vasculitis. Therapy may be restricted to withdrawal of the most likely offending drug. Refractory cases have been managed with prednisone (0.5 mg/kg per day), colchicine (0.6 mg twice daily), dapsone (150 mg/d), and rarely cytotoxic agents or azathioprine. Topical therapy adds little in the absence of skin breakdown. Complete resolution may take weeks.

Hepatitis C Virus Infection

A significant number of patients with hepatitis C virus (HCV) infection can develop urticaria, livedo reticularis, leukocytoclastic vasculitis, and membranoproliferative glomerulonephritis. The syndrome is mediated by HCV-driven formation of cryoglobulins that act as immune complexes. Indeed, it is highly likely that almost all patients previously diagnosed as idiopathic mixed cryoglobulinemia with vasculitis were actually patients with HCV. High titers of rheumatoid factors and hypocomplementemia strongly support the diagnosis. Treatment is directed at control of viral replication in the liver and attenuation of immune complex activity by plasmapheresis and cyclophosphamide. Until the role of B cells in defense against HCV is better understood, the temptation to deplete antibody-producing cells with anti-CD20 monoclonals should be approached with caution.

Ischemic Digits: Is This Vasculitis?

Patients in the ICU may develop ischemic digits. Contributing factors include hypotension, use of radial arterial lines, and vasoconstrictors. Often the issue of vasculitis is raised.

Single extremity involvement speaks strongly against systemic vasculitis as the cause. Similarly, isolated toe involvement is more likely to be caused by a combination of noninflammatory vascular disease and diminished blood flow owing to hypotension, vasoconstrictors, or cholesterol emboli. The latter may shower from the aorta and create a pseudovasculitic picture, particularly after anticoagulation therapy is initiated or following instrumentation of the aorta.53 Extreme symmetry of lesions with all digits involved is more suggestive of a generalized low-flow state than vasculitis.

Clues to vasculitis as the cause of digital ischemia include the coexistence of a disease associated with digital vasculitis such as SLE or scleroderma, random involvement of multiple limbs, the presence of nailbed infarcts, and other associated cutaneous markers specific to vasculitis such as palpable purpura. Likewise, extracutaneous markers of vasculitis including glomerulitis and patchy neurologic deficits would enhance suspicion for that diagnosis.

Male patients with a history of heavy smoking should be suspected of Buerger's disease (thromboangiitis obliterans). Patients with lupus and other connective tissue diseases who have ischemic digits may have thrombotic complications secondary to ACL antibody and not true vasculitis.54

Biopsy of ischemic digits is usually impractical and potentially hazardous. The necessity to amputate a gangrenous digit should prompt careful instruction to the surgeon to be sure to biopsy the digital artery immediately proximal to the gangrene. In this setting angiography often reveals nonspecific findings of small-vessel disease, but may suggest emboli, and at times reveal the source such as a subclavian plaque. An angiographic pattern suggestive of Buerger's disease has been described.55

Lung Infiltrates and Elevated Sedimentation Rate: Is This Vasculitis?

Elevated ESRs frequently become laboratory aberrations looking for a disease state. The unwary clinician may let an elevated ESR be the driver for a costly and unrewarding work-up. Once it is known that a patient has an elevated ESR and the usual causes for this nonspecific laboratory abnormality have been reasonably excluded, the impulse is to link any and all remaining clinical abnormalities to the abnormal ESR. Unexplained pulmonary infiltrates, a common finding in critically ill patients, are good examples that provoke the question: Is this a pulmonary vasculitis? Common sense should prevail and dictate that the diagnosis of pulmonary vasculitis as a principal entity should be one of exclusion. If a young female on corticosteroids with a fever, rash, alopecia, and pericarditis develops a pulmonary infiltrate, she should be considered to have an infectious disease or pulmonary embolism until otherwise proved. If a previously bedridden patient with advanced rheumatoid arthritis suddenly develops pulmonary infiltrates and hypoxemia, the etiology should be considered embolic until proved otherwise. If a patient with advanced scleroderma that includes the proximal gut develops pulmonary infiltrates, aspiration pneumonia should be strongly considered. Conversely, one can approach the patient with pulmonary infiltrates and an elevated ESR in reductionist fashion.

If a careful search for extrapulmonary evidence of vasculitis is not rewarding, it will be highly unlikely that primary pulmonary vasculitis is present, since vasculitis very rarely involves only the pulmonary tree. There are unusual instances when vasculitis can be isolated to the pulmonary tree in a systemic rheumatic disease. Rarely, patients with ACL antibodies may develop pulmonary infiltrates secondary to in situ pulmonary artery thrombosis and/or pulmonary embolism. These patients will simulate a primary vasculitic pulmonary picture, but the primary therapy is anticoagulation. The proclivity of lung infections (especially bacterial) to generate positive anti-neutrophil cytoplasmic antibodies (ANCA) can create a diagnostic quandary when the question of Wegener's granulomatosis is raised.

Erythrocyte Sedimentation Rate

The ESR is an indirect determination of the acute phase response and may be elevated in the setting of infection or active rheumatic disease. Values are higher for women and the elderly. Although the exact appropriate adjustments for age and sex are not certain, a common rule used to calculate a “normal” ESR in patients over 40 is divide the age by 2 and add 10 if the patient is female.

The presence of monoclonal proteins, polyclonal hypergammaglobulinemia, hyperfibrinogenemia, and alterations in size, shape, and number of red blood cells will influence the ESR. ESR rises during normal pregnancy and should not be used to monitor rheumatic diseases under these circumstances. The ESR increases in end-stage renal failure of whatever cause and is not indicative of an underlying rheumatic disorder.56 Parallel confirmation of a normal C-reactive protein level often clarifies the noninflammatory origin of the elevated ESR. The level of the rise in ESR correlates imperfectly with disease activity and may at times be normal in patients with active rheumatoid arthritis or SLE. Patients with a markedly elevated ESR (MESR) are those with values >100 mm/h. These patients deserve special attention because such elevations are unlikely to be explained by age or normal physiologic state and are a more reliable sickness indicator. The illnesses associated with MESR include infection, malignancy, rheumatic disorders such as vasculitis, connective tissue disease, rheumatoid arthritis, and temporal arteritis, as well as end-stage renal failure, nephrotic syndrome, and other inflammatory diseases such as hepatitis and colitis. In most series looking at MESR, 3% to 10% of patients will have no diagnosis to explain the abnormal laboratory value.57 Some of these patients will eventually reveal an underlying pathology, whereas others will demonstrate spontaneous improvement in the ESR.

In the ICU, the ESR is likely to be elevated for multiple reasons. An MESR should not prompt an unreasonable search for vasculitis or other concurrent rheumatic disease, particularly in the presence of renal failure and nephrotic syndrome.

C-Reactive Protein

The availability of rapid, reproducible, high-sensitivity, low-cost assays for C-reactive protein (CRP) in many circumstances obviate a clinician's reliance on the vagaries of the ESR to detect inflammation. CRP is an acute phase serum protein that acts as a surrogate for the proinflammatory interleukin IL-6. Serum/plasma levels of CRP are tightly linked not only to the absolute level of IL-6, but also to the rise and fall of the cytokine. Sequential CRP determinations, in contrast to the slowly responsive ESR, can provide a more accurate assessment of inflammatory changes in response to therapy. CRP determination is most helpful in assigning a noninflammatory cause to a markedly abnormal ESR. An example would be a patient with a monoclonal protein without infection, who could have an ESR of 100 but a normal CRP. The latter alleviates concern that significant inflammation may be part of the clinical picture.

Several cautions apply to the interpretation of CRP levels. Although there are weak correlations between CRP levels and age and gender, they are less marked than the same variables and the ESR. CRP also varies directly with body mass index. The very new ultrasensitive CRP assays have not only changed the normal ranges, but are expressed in markedly different concentrations that can be very confusing to the unwary.

The bottom line for the critical care clinician might be that significant elevations of CRP usually signal the presence of clinically relevant inflammation, and the absence of a high CRP helps in excluding it.

Antinuclear Antibodies

The presence of antinuclear antibodies (ANA) in high titer provides presumptive evidence for the presence of collagen vascular disease, and in particular, SLE (Table 104-1). Lower levels of ANA can be nonspecific, possibly normal, and at times explained by age, prior drug therapy, or a first-degree relative with lupus. Low levels of ANA in the elderly can be particularly misleading in the presence of an age-related elevation of the ESR. It has been estimated that 15% to 25% of normal, healthy individuals over the age of 60 will have circulating ANA.58 These data are confounded by a small but definite incidence of new cases of SLE among the elderly.

Screening ANA is typically done by a standard, indirect immunofluorescence technique. The test is now performed most commonly on HEp-2 cells and is sensitive for detecting the presence of SLE and other collagen vascular diseases, but as noted above, the test is hindered by the lack of specificity. Interest in specific ANA has spawned a long and at times confusing list of tests of variable utility. A brief overview of the most useful specific ANA tests follows.

Anticentromere Antibody

This antibody to the kinetechore of chromosomes is detected by recognition of a particular speckled pattern of immunofluorescence on HEp-2 cells. It is, in general, the only pattern detected on screening ANA useful for diagnostic purposes. The antibody is found most commonly in the limited cutaneous variant of scleroderma (CREST). In this subset of patients, the test has been positive in 44% to 98% of those tested.59,60 Less commonly, it may be seen in diffuse scleroderma and primary biliary cirrhosis with or without evidence of scleroderma.

Antibodies to DNA

Antibodies to DNA fall into two major categories by virtue of reacting to antigenic determinants on the phosphate deoxyribose backbone of the DNA helix or determinants on the nucleotide bases. The former represent antibodies to native double-stranded DNA while the latter react with single-stranded DNA.

Antibodies to single-stranded DNA are more common and are found across a spectrum of rheumatic and nonrheumatic disorders. They are of no practical clinical utility. Antibodies to double-stranded DNA are useful since they have great specificity for SLE and are found in 60% to 70% of patients with that disease.43,61 In many, but not all lupus patients, levels of anti-DNA antibody will correlate positively with disease activity. Low levels of this antibody have occasionally been found in other connective tissue diseases.

Antibodies to Sm

This antibody is named after a patient (“Smith”) in whom it was first described. The antibody has high specificity for SLE and is rarely found in patients with other connective tissue diseases. Sensitivity is only about 30% for SLE. Sm is not to be confused with an antibody to smooth muscle (SM), which is not a marker for collagen vascular disease, but is found in patients with chronic liver disease. There is no specific clinical profile of Sm-positive patients with SLE. Titers are not useful for assessment of disease activity.

Antibodies to nRNP

Antigenic determinants for nuclear ribonucleoprotein (nRNP) may occur in a molecular complex with Sm, and antibodies to Sm and RNP are often found in the same patient. Antibodies to nRNP may be seen in SLE, scleroderma, or overlap syndromes. The presence of overlapping clinical features and high titers of antibody to RNP defines a clinical subset of patients referred to as those with mixed connective tissue disease (MCTD).

Antibodies to SS-A/Ro and SS-B/La

These antigens were originally described in patients with Sjögren's syndrome (SS) and SLE. They are RNA-protein conjugates. SS-A and Ro have antigenic identity, as do SS-B and La. The Ro and La refer to the antigen as localized to the cytoplasm, whereas SS-A and SS-B are nuclear antigens.

The presence of SS-B/La may be seen in SLE or Sjögren's syndrome, and in most assays is measured along with SS-A/Ro. The Ro antibody has been described in 60% of so-called ANA-negative SLE. Anti-Ro antibody is also highly prevalent in the setting of congenital heart block and neonatal SLE.62 In those cases, the antibody is found in mother and child. Other clinical scenarios associated with anti-Ro antibody include subacute cutaneous lupus and C2 deficiency. Anti-Ro antibody occurs in 25% to 40% of unselected patients with SLE.

The major indications for ordering these tests are: in a setting in which SLE is strongly suspected but the screening ANA is negative, congenital heart block, neonatal lupus, and the initial evaluation of a patient with a positive ANA.

Antibodies to Scl-70 (Topoisomerase I)

Antibodies to Scl-70 are directed toward DNA topoisomerase I and inhibit its function.63 They are found in 20% to 40% of patients classified as diffuse systemic sclerosis, and less commonly in patients with limited cutaneous disease or CREST. Determination of this antibody is part of the evaluation of patients suspected of having scleroderma.

Antineutrophil Cytoplasmic Autoantibodies

The detection of antibodies directed against neutrophil cytoplasmic components now offers a useful serologic tool for the diagnosis and management of a group of disorders characterized by systemic necrotizing vasculitis and glomerulonephritis.64,65 These disorders include Wegener's granulomatosis, microscopic polyangiitis, Churg-Strauss syndrome, and idiopathic crescentic glomerulonephritis. The renal lesions in these disorders have in common necrotizing vascular injury and a paucity of immune deposits.

Antineutrophil cytoplasmic autoantibodies (ANCA) are found in 90% of patients with active, generalized Wegener's granulomatosis and 60% to 70% of those with limited disease. The titer often parallels disease activity and may be helpful in distinguishing a disease flare from intercurrent infection or other morbidity in patients with Wegener's granulomatosis. ANCA can be found in 80% of patients with active pauci-immune necrotizing and crescentic glomerulonephritis, which is one of the major causes of rapidly progressing glomerulonephritis.

Two specific patterns of ANCA have been identified and are referred to as cytoplasmic ANCA (C-ANCA) and perinuclear ANCA (P-ANCA). In general, patients with Wegener's granulomatosis have demonstrated C-ANCA, whereas those with idiopathic crescentic glomerulonephritis have demonstrated P-ANCA.66 P-ANCA has been reported with other forms of systemic vasculitis, including microscopic polyangiitis and Churg-Strauss vasculitis. P-ANCA has also been observed in patients with ulcerative colitis and Crohn's disease.

The clinical scenarios that warrant measurement of ANCA are patients with known or suspected Wegener's granulomatosis or other small-vessel vasculitis and patients with rapidly progressing glomerulonephritis. These antibodies have also been occasionally identified in Takayasu's disease, SLE, relapsing polychondritis, and Behçet's disease.

Antibodies in Polymyositis/Dermatomyositis

In general, serologic testing has been of little practical value in the diagnosis and management of patients with polymyositis or dermatomyositis. Recently, the emergence of a family of autoantibodies that are found nearly exclusively in patients with myositis, known as myositis-specific autoantibodies or MSA, has focused interest on the role of humoral immunity in this disease. MSA are usually directed against intracellular, intracytoplasmic antigens involved in protein synthesis. Specific ANA have been identified in patients with these diseases and are available commercially. The exact role of these antibodies in the management of patients is unclear, but identification of these antibodies may be useful in patients who represent a diagnostic dilemma.

Jo-1 antibody is the most common of the MSA and belongs to a family of autoantibodies known as antisynthetases. It can be found in 15% to 40% of patients with polymyositis, and less commonly in dermatomyositis. The presence of the antibody highly correlates with associated interstitial lung disease. Antibodies to the nuclear antigen Mi occur in about 5% of patients with dermatomyositis and rarely in polymyositis. MSA titers have not proved useful in monitoring the course of patients with polymyositis or dermatomyositis. Specific MSA may prove to define unique clinical and immunogenetic groups that represent separate but related diseases. Antibody to PM-1 or PM-Scl defines a small subset of polymyositis patients (10%), half of whom will have accompanying features of scleroderma. Occasionally, patients with scleroderma and no myositis will have this antibody.

Rheumatoid Factor

Rheumatoid factors (RFs) are autoantibodies, predominantly IgM isotype, that are directed against multispecies antigenic determinants on the heavy chain of IgG. They are nonspecific, rarely diagnostic, and almost never useful in the critical care setting. These autoantibodies can be found in normals and many disease states and have a prevalence that varies directly with age and gender (female preponderance). RFs can arise as an epiphenomenon during acute illness or chronic antigenic stimulation of almost any cause. They are present, sometimes in significant titer, in bacterial endocarditis, granulomatous diseases, and most rheumatic diseases at some point in time. The presence of RFs, occurring in tandem with significant decreases of serum complement components C3 and C4, may provide a diagnostic clue in rarely encountered clinical syndromes such as rheumatoid vasculitis with cryoglobulins or hepatitis C–associated mixed cryoglobulinemia and vasculitis. These syndromes can present with gastrointestinal tract involvement and hemorrhage, compromised renal function or progressive peripheral neuropathy, and skin ulceration. The recognition of these vasculitic syndromes mandates aggressive intervention with cytoreductive therapy—usually cyclophosphamide and therapeutic plasmapheresis.67,68

Complement Levels

In theory, the detection of direct (classic) complement pathway activation by the measurement of serum C3 and C4 levels should provide the clinician with laboratory support for using treatment protocols that suppress the sequelae of complement activation. A prototype disease where C3 and C4 changes should be helpful is SLE. Unfortunately, there is an imperfect correlation between clinical activity of any disease, especially SLE, and decreased C3 and C4 levels. In fact, reliance upon conventional measurements of serum C3 and C4 assays, which supposedly detect immune complex complement activation in the patient, can inappropriately drive therapy decisions toward the use of immunosuppressives or corticosteroids when they are not indicated.69

Moreover, C3 synthesis is increased by acute inflammation of any cause, serving as a classic acute-phase reactant. It is easy to visualize a scenario in which a patient with SLE being treated with corticosteroids who has increased consumption of C3 secondary to immune complex formation and deposition may develop secondary bacterial infection, which stimulates the production of C3 in its role as an acute-phase reactant. The end result will be a normal serum level of C3, which may engender a false sense of security with regard to SLE disease activity. Conversely, the significantly reduced synthesis of C3 in many hepatic diseases is reflected in low plasma levels that can be misinterpreted. Plasma C4 is a more sensitive indicator of direct pathway activation, but for many of the same reasons that relate to C3 determinations, a decreased C4 level should never be the sole determinant of intervention with potent immunosuppressive therapy. Patients with SLE or overlap variants may have heterozygous or homozygous defects in C4 production. These patients will always have low C4 levels, regardless of disease activity.

ICU patients admitted with either meningococcemia or gonoccocemia provide one of the rare reasons for the determination of total hemolytic complement level. This assay serves as a screening test that depends on the functional presence of all individual complement components. A significant decrease in hemolytic activity may identify patients with terminal complement component deficiency who are at high risk for recurrent bacteremia. Such patients may require fresh frozen plasma in conjunction with antibiotic therapy and should be immunized with meningococcal vaccine prior to discharge.

In summary, conventional measurements of serum complement levels are not helpful in predicting activation in this complex system. When sensitive assays for activation peptides or complement activation complexes become available, assessment of the complement pathway may become more clinically relevant.

Cryoglobulins

Cryoglobulins are monoclonal, oligoclonal, or polyclonal immunoglobulins with a thermal propensity for precipitation at decreased temperatures and subsequent resolubilization with warming. In the vast majority of instances, they occur as acute or chronic phase reactants during bacterial or viral infections and many rheumatic disease syndromes. Plasma cryoglobulins, especially when assessed as cryocrits, are very nonspecific and may simply reflect ongoing antibody formation and subsequent complexing with antigen; indeed, they have been termed “the poor man's immune complex assay.” Abnormal elevations of the ESR can be masked by some cryoglobulins that form gels at ambient temperatures and impede the rate of erythrocyte descent. Pseudoleukocytosis can occur when automated cell-counting procedures count crystallized cryoproteins as white blood cells. Pseudohypogammaglobulinemia and pseudothrombocytosis have also been reported. Almost all patients with cryoglobulin-mediated immune complex syndromes have active HCV infection that should be aggressively treated. Cryoglobulin-mediated vasculitis should be followed in terms of its clinical activity; clinicians should not rely on cryocrit decreases during therapy. Falls in cryocrits may be misleading because cryoproteins, even at extremely low levels in plasma, can be highly efficient complement activators.

Anticardiolipin Antibodies

The anticardiolipin (ACL) antibodies belong to a family of antiphospholipid antibodies (APLA), including those responsible for the lupus anticoagulant (LA), the biologic false-positive test for syphilis, and anti-B2-glycoprotein-I antibodies. These antibodies often, although not always, occur together. The clinical syndromes associated with these antibodies belong to a growing list that can be explained largely by the capacity of these antibodies to induce thrombosis in the venous and arterial circulation. Thrombocytopenia and recurrent fetal loss are the other major consequences of APLA. The combination of APLA and one or more of these clinical features has been termed the “antiphospholipid antibody syndrome.”70

Chronic false-positive serologic tests for syphilis are associated with autoimmune disease, notably SLE, and are found with increased frequency in patients with ACL and LA activity. Lupus anticoagulants are IgG or IgM antibodies that prolong phospholipid-dependent tests in vitro by interference with the calcium-dependent binding of prothrombin (factor II) and factor Xa to phospholipids, thus inhibiting the generation of prothrombinase. This usually results in prolongation of the activated partial thromboplastin time (APTT) with or without slight prolongation of the prothrombin time (PT).54 LA is a common cause of prolongation of the PTT, but not the only cause. Most patients with LA do not have SLE.

Usually ACL antibodies are detected in an enzyme-linked immunosorbent assay (ELISA) using bovine cardiolipin as substrate. These are the most commonly detected antiphospholipid antibodies. They are generated transiently in the course of acute infections including mycoplasma and gram-negative infections. These antibodies are usually of the IgM isotype and are not associated with thrombosis. IgM ACL as well as the LA may be induced by a variety of drugs including phenothiazines, procainamide, phenytoin, hydralazine, quinidine, and streptomycin. These antibodies are most often not associated with thrombotic events, but exceptions to this rule occur. Antibodies to B2-glycoprotein-1 in the absence of phospholipids can be measured on oxidized polystyrene plates. These antibodies are more specific for the diagnosis of APLA syndrome than the anticardiolipin test.

ACL antibodies are noted in 2.5% of the general population.54 For most of these patients, the antibodies have no clinical significance. The risk of thrombosis and fetal loss has been generally associated with higher levels of antibody and the IgG isotype. Unfortunately, exceptions occur. Thus the presence of ACL antibody should not in itself prompt therapeutic intervention.

A myriad of neurologic events including stroke, transient ischemic attacks, and amaurosis fugax have been associated with the presence of LA and ACL antibody. Their presence should be suspected in patients who have no risk factors for thrombosis or who have associated autoimmune disease or suggestive screening laboratory abnormalities, including prolonged PTT or false-positive serologic tests for syphilis. Skin lesions secondary to LA and ACL include gangrene, hemorrhage, and purpura. Necrosis may mimic vasculitis.

The vasculopathy of APLA syndrome is not vasculitis, but primarily thrombosis of large or small arteries or veins. Steroids may normalize the LA but have little impact on ACL antibody levels. Treatment for major thrombotic complications of this syndrome is heparin or enoxaparin and warfarin. Steroids are indicated for associated clinical features related to systemic inflammation.

Rarely patients may present with evidence of widespread vascular occlusion and multiorgan failure occurring concurrently or over a short period of time related to antiphospholipid antibodies. Kidneys, bowels, lungs, brain, and heart are frequently involved. This so-called catastrophic antiphospholipid syndrome (CAPS) is associated with significant morbidity and mortality in spite of empiric therapy with corticosteroids, anticoagulation, and apheresis.72 CAPS needs to be distinguished from thrombotic thrombocytopenia purpura (TTP) and diffuse intravascular coagulation. Precipitating events, such as infection, trauma, surgical procedures, or reduction in anticoagulation therapy, may contribute to the development of CAPS.73

Corticosteroids

Corticosteroids have potent immunosuppressive and anti-inflammatory properties that, in combination with rapid onset of action, make them the drugs of choice for the initial therapy of most acute, life-threatening rheumatic disorders. Even low doses of prednisone (<10 mg/d) have potent anti-inflammatory effects and are highly effective in patients with rheumatoid arthritis.

For purposes of controlling inflammation and immunomodulation, “short-acting” glucocorticoids with little or no mineralocorticoid activity are preferred. The oral drug of choice is prednisone, which is converted to prednisolone in the liver. Whereas active liver disease impairs that conversion, it appears to be offset sufficiently by decreased rate of elimination of prednisolone to eliminate the need to preferentially use prednisolone in patients with cirrhosis or active liver disease. The intravenous drug of choice is methylprednisolone. The dose equivalency is 4 mg methylprednisolone to 5 mg prednisone.

The dose of prednisone or methylprednisolone is largely empiric. For serious, life-threatening problems, 1 mg/kg per day of prednisone is a good starting point. Dividing the dose into a twice-a-day or other dose-divided schedule may increase efficacy (as well as toxicity) and is recommended by some for initial therapy. Extremely large doses of intravenous methylprednisolone (500 to 1000 mg) daily have been used for brief periods (3 to 5 days) with variable success in a variety of clinical settings, mostly in the context of SLE. In vitro studies suggest that such large doses may produce a qualitatively different response in lymphocyte function.74 Unique side effects to this form of therapy, including sudden overwhelming sepsis and sudden death, are rare. This form of therapy is generally reserved for patients who have failed conventional high-dose therapy with corticosteroids with or without another immunosuppressive agent.

Patient response to corticosteroids varies. Failure to respond is likely a result of the nature and severity of the disease. The effectiveness of glucocorticoids may be reduced by simultaneous use of other drugs that induce hepatic microsomal enzyme activity, such as phenytoin, barbiturates, and rifampin. Bioavailability of prednisone may be reduced by antacids sometimes prescribed for concurrent use. Cortisol and its synthetic derivatives are bound to corticosteroid-binding globulin and albumin. The bound steroid is not active. Increased frequency of prednisone side effects has been observed at low serum albumin levels, probably reflecting an increase in the unbound, active fraction of the drug.

Patients who have a positive purified protein derivative (PPD) test about to undergo corticosteroid therapy (particularly with doses of prednisone of 20 mg/d or greater) should be considered for isoniazid (INH) prophylaxis (300 mg/d orally). The reported risk of reactivation ranges from low in asthmatics to higher in the elderly and in patients immunosuppressed by virtue of other drugs or their primary disease. The patient with a positive PPD and either a normal chest x-ray or a single calcified nodule probably does not require prophylaxis.76 If the patient has significant impairment of the immune system or the chest film shows fibronodular scarring, the risk is enhanced considerably, and prophylaxis with INH is advisable.77

Steroid therapy suppresses cutaneous delayed hypersensitivity responses by inhibiting recruitment of macrophages to the skin test site. This phenomenon is reversible on stopping the drug. In one study, treatment with 10 mg prednisone daily totally inhibited cutaneous tuberculin sensitivity in both active and inactive cases of tuberculosis, with a mean reversion time of 13.6 days and reconversion time of 6 days following discontinuation of the drug.

Acute adrenocortical insufficiency may occur in critically ill patients who have been treated with chronic glucocorticoid therapy. On the basis of available data, any patient who has received a glucocorticoid at a dose equivalent to 20 to 30 mg prednisone daily should be suspected of having hypothalamic-pituitary-adrenal (HPA) axis suppression.78 At doses closer to but above the physiologic range, a month is probably the minimum duration required for HPA suppression. Patients receiving the equivalent replacement doses of steroid (5 mg prednisone) as single morning dose therapy are at low risk of iatrogenic adrenal insufficiency. In the absence of hemodynamic instability, these patients do not require full “stress dose” replacement therapy. An adrenocorticotropic hormone stimulation test can resolve the question of adrenal suppression, but the clinical reality usually dictates empirical coverage with “stress doses” of corticosteroids. This can be accomplished with 100 mg hydrocortisone intravenously every 8 hours. This is approximately the equivalent of 75 mg prednisone. Higher doses are not necessary and are potentially more hazardous.

Use of corticosteroids on alternate days is associated with some decrease in chronic drug morbidity. In the urgent setting of the critically ill patient, the greater effectiveness of daily or split daily doses of steroids recommends such a dosing schedule.

Immunosuppressives

Cyclophosphamide

Cyclophosphamide is an alkylating agent with broad immunosuppressive properties. It is generally embraced as the drug of choice for suppression of progressive, life-threatening autoimmune disease unresponsive to corticosteroids alone. Effectiveness has been reported in a broad spectrum of collagen vascular diseases and primary vasculitides, including SLE, Wegener's granulomatosis, polyarteritis nodosa, and polymyositis/dermatomyositis.

Cyclophosphamide causes broad suppression of B- and T-cell function and acts as a potent inhibitor of antibody production. Anti-inflammatory effects have also been described.

The drug is rapidly absorbed orally. It is inert until metabolized in the liver. Extravasation of the drug is not caustic to soft tissues. Sixty percent of the drug is excreted in the urine in the form of active metabolites. Impaired excretion of these active metabolites because of renal insufficiency can potentiate the therapeutic and toxic effects of a given dose of drug.

The drug can be given orally, usually at a dose of 2 mg/kg per day, or intravenously. To circumvent toxic effects associated with chronic drug exposure, intravenous bolus cyclophosphamide therapy has become an increasingly popular alternative at doses of 0.5 to 1.0 g/m2. The onset of immunosuppressive activity of cyclophosphamide is estimated at 10 to 14 days following initiation of therapy. Although unproved in rigorous clinical trials, there is an operational principle that immunosuppression can be achieved more rapidly with intravenous bolus therapy. Hence bolus cyclophosphamide is most often given in the setting of progressive life-threatening disease requiring immunosuppression. A major short-term side effect of bolus therapy is a predictable white blood cell count nadir 7 to 10 days after drug infusion. Leukocyte count levels typically recover in 2 to 3 days. However, if the patient has a concurrent bacterial infection during the nadir period, the consequences of even transient profound neutropenia can be disastrous. Gross hematuria may signal the development of hemorrhagic cystitis or bladder malignancy. Bladder problems are related to duration of therapy and total cumulative dose administered. Many of the other side effects of therapy with cyclophosphamide are related to chronic use, including gonadal suppression, oncogenesis, pulmonary interstitial fibrosis, and hypogammaglobulinemia.

One report suggests that bolus cyclophosphamide is less effective than oral cyclophosphamide in patients with Wegener's granulomatosis.71 This study raises questions about the predicted superiority of bolus cyclophosphamide in all clinical situations.

Azathioprine

Azathioprine is a commonly used immunomodulating drug with mild to moderate immunosuppressive properties that may in large part be explained by a preferential reduction of natural killer cells. Onset of action is slow, probably taking months. It is often used concurrently with corticosteroids in patients requiring unacceptably high doses of steroids, to reduce the steroid dose. Azathioprine is not the drug of choice when significant immunosuppressive effect is needed on an urgent basis. Risk for infection is modest in the absence of leukopenia.

Azathioprine is metabolized in the liver to the active metabolite, 6-mercaptopurine, a purine analogue. The drug interferes with purine biosynthesis and is ultimately metabolized by xanthine oxidase. Administration of allopurinol, which inhibits xanthine oxidase, to a patient on a stable dose of azathioprine may result in a fatal drug overdose. The azathioprine dose should be reduced by 50% to 75% in the presence of allopurinol.

Dose range for this drug is 1 to 3 mg/kg per day. The drug is primarily metabolized in the liver, but the need for dose adjustment in the presence of liver disease is variable and may be unnecessary. Drug half-life can increase in renal failure but may not prove clinically significant. Cautious observation for the development of cytopenia is indicated in the presence of hepatic or renal failure. The drug is well absorbed orally and may be given intravenously in doses equivalent to the oral form.

Mycophenolate Mofetil

Mycophenolate mofetil (MMF) interferes with de novo purine synthesis by inhibiting inosine monophosphate dehydrogenase. Nonlymphocytic cells recycle purine precursors through salvage pathways and are not dependent on de novo purine synthesis, in contrast to proliferating T and B lymphocytes, which are completely dependent on the de novo pathway. This difference confers selectivity on MMF and makes it an attractive choice for preventing allograft rejection. MMF is rapidly supplanting azathioprine in multiagent immunosuppressive protocols used in heart, lung, and renal transplantation. The conventional daily dose of MMF is usually 2.0 g and greater therapeutic effects are not usually found with higher doses. Mild reversible leukopenia and mild gastrointestinal side effects have occurred with MMF, and dose adjustments should be made when a patient has a glomerular filtration rate less than 25% of predicted value. There is a low incidence of opportunistic infection that can be directly attributed to MMF when used at less than 3 g daily. In most clinical scenarios, plasma levels of its active metabolite, mycophenolic acid, do not need to be monitored to achieve therapeutic effects. Based on its selectivity, ease of administration, and low side-effect profile, there is increasing interest in using MMF for the treatment of patients with autoimmune diseases.

Methotrexate

Although methotrexate was introduced for the treatment of rheumatoid arthritis more than 30 years ago, several open trials in the early 1980s resulted in renewed interest in this drug. Low-dose oral methotrexate has now replaced gold as the remitting agent of first choice in rheumatoid arthritis for many rheumatologists. The drug is accepted as effective in psoriatic arthritis and Reiter's syndrome as well as polymyositis. Experience is being gained with its use in SLE and scleroderma. Use of methotrexate in the critically ill rheumatic disease patient is currently limited to patients with polymyositis/dermatomyositis that is refractory to corticosteroids.

Methotrexate is a folic acid analogue and the major folic acid antagonist in clinical use. The drug is absorbed after oral ingestion but with significant variability. More predictable serum levels can be achieved by intramuscular or intravenous administration. High-dose methotrexate can alter antibody production and cellular immunity. Low-dose oral methotrexate (25 mg/wk or less), as used in rheumatoid arthritis, may be mainly anti-inflammatory or directly inhibiting to synovial lining cells.

Low-dose methotrexate is given in rheumatoid arthritis in initial doses of 5.0 to 10 mg and may be gradually increased to levels of 25 mg/wk. Methotrexate for rheumatic disorders is delivered on a weekly basis. This regimen is associated with less toxicity than when the drug is given more frequently, particularly hepatotoxicity. Most patients will respond at doses between 10 and 20 mg weekly. Intravenous methotrexate for myositis is given in doses ranging from 25 to 75 mg weekly. A dose of 1 mg/kg per week has been used in childhood dermatomyositis/polymyositis.79 Such a dose calculation may be a reasonable first approach in the treatment of adults.

Adverse reactions forcing discontinuation of the drug in short-term trials with rheumatoid arthritis occur in 5% to 31% of patients. Most toxicity is relatively minor and associated with advanced age, malnutrition, and impaired renal function. Nausea, vomiting, oral ulcers, rash, leukopenia, thrombocytopenia, and pancytopenia all may occur. Cirrhosis may occur in some patients treated for long periods and appears to be related to cumulative dose and probably the nature of the underlying disease being treated. The risk is asserted to be greater in psoriatic arthritis than rheumatoid arthritis. Ethanol potentiates the hepatotoxicity of methotrexate. Pretreatment screening for hepatitis B and C is usually done. Baseline liver biopsy is not indicated in the absence of risk factors for existing liver disease. The concurrent use of other antifolates (e.g., sulfonamides) increases toxicity. Pancytopenia has been reported in some patients receiving methotrexate and trimethoprim-sulfamethoxazole together. Patients with ascites and large effusions are at greater risk for methotrexate toxicity. Folic acid, 1 mg daily, has been recommended as a means of preventing adverse reactions and particularly hematologic side effects in patients treated for rheumatoid arthritis.80 The use of folic acid does not appear to significantly reduce effectiveness of the drug. In serious episodes of pancytopenia, leucovorin may be used.

Toxicities of particular concern in ICU patients include leukocytoclastic vasculitis, which has been reported with high-dose therapy used to treat osteogenic sarcoma as well as low-dose therapy for rheumatoid arthritis. Opportunistic infections with herpes zoster and Pneumocystis carinii have been reported even with low-dose methotrexate, although they are uncommon. Finally, an acute hypersensitivity pneumonitis as discussed earlier occurs infrequently but may result in profound hypoxemia.

Calcineurin Inhibitors

The emergency use of cyclosporine and tacrolimus outside the setting of clinical transplantation will be rare. Their role in the urgent therapy for fulminant psoriasis and inflammatory bowel disease has been superseded by the availability of anti–tumor necrosis factor (TNF)-α biopharmaceuticals. The mechanism of action and dosing intricacies of both drugs are very similar. There is an accelerating trend favoring the use of tacrolimus in organ transplantation because the incidence of hypertension, hyperlipidemia, and nephrotoxicity appear to be less than with cyclosporine. If rapid achievement of therapeutic blood levels is needed, continuous intravenous administration of the calculated daily dose is required. The intravenous route is mandatory in children and patients with hypermotile or maladsorptive gastrointestinal disease. In a critically ill patient, maintenance of therapeutic plasma levels of a calcineurin inhibitor requires intensive monitoring of trough drug levels and frequent dose adjustments. A disparate group of commonly used drugs affect calcineurin inhibitor metabolism and eternal vigilance is required to prevent sudden decreases to subtherapeutic levels or increases that are nephrotoxic. The volume-depleted patient is especially susceptible to renal failure caused by either calcineurin inhibitor. Acute, severe gouty arthropathy is extremely common in patients being treated with cyclosporine.

High-Dose Intravenous Globulin

Intravenous infusion of pooled IgG antibodies from large numbers of normal volunteers activates a diverse spectrum of immunomodulatory effects. The primary mechanism may be mediated by downregulation of Fc receptors on neutrophils and macrophages and inhibition of B-cell antigen receptors, but cytokine, idiotypic, and major histocompatibility antibodies may play a role also.86 The spectrum of immunomodulation provides a rationale for the use of intravenous immunoglobulin (IVIg) in the treatment of life-threatening vasculitic syndromes of unknown etiology. IVIg can also be useful in patients who require immunosuppression or modulation, but cannot tolerate cyclophosphamide because their bone marrow is suppressed or hypoplastic.

High-dose intravenous antibody therapy may be indicated for therapy-resistant immune-mediated thrombocytopenia associated with significant bleeding, especially gastrointestinal bleeding, or when there is a need to either transiently elevate platelets or prolong the half-life of transfused platelets prior to splenectomy.75 These modified, biologically active immunoglobulins are given at approximately a 0.4-g/kg dose and followed by platelet transfusions as indicated. Imaginative uses of these expensive biologics are frequent, usually in disease settings where all else has failed. Appropriate clinical trials are needed before their widespread use. There are few safety issues with the use of IVIg, but several instances of large vein thrombosis at the site of IVIg infusion have been reported.

The theoretical risk of generating anaphylatoxins or other complement-derived vasoactive mediators during rapid infusion of modified immunoglobulin preparations has been overstated.

Anti-Inflammatory Agents

Nonsteroidal Anti-Inflammatory Drugs

The NSAIDs generally have little role to play in the management of patients in the ICU. Ulcerogenesis, hemostatic defects, and impairment of renal function converge as a group of notably unfriendly toxic effects for the critically ill patient. In general, corticosteroids are usually a safer alternative when an anti-inflammatory agent is necessary in the ICU. For instance, corticosteroids in low doses can be used to control symptomatic synovitis in the rheumatoid patient and are considerably less hazardous than NSAIDs. Moderate-dose oral, parenteral, or intra-articular steroids may be used for gout.

The emergence of drugs that selectively block the cyclooxygenase isoform COX-2, an inducible isoenzyme involved in inflammation, has created a safer class of NSAID. Selective COX-2 inhibitors including celecoxib, rofecoxib, and valdecoxib have been associated with a reduction in major gastrointestinal complications when compared to traditional NSAIDs. They are not renal sparing and do not appear to be safe in patients with aspirin-sensitive asthma. Nonetheless, because of the absence of platelet effect and lower GI toxicity they would be preferred agents if an NSAID was required. The concurrent use of a proton pump inhibitor to further reduce the risk of GI toxicity would seem prudent in the critically ill patient.

COX-2 drugs have recently been implicated in an increased risk of myocardial infarction and stroke. The story is unfolding as this book goes to press.

Colchicine

Colchicine is a plant-derived alkaloid that exerts a complex array of physiologic changes in the neutrophil. This appears to explain its therapeutic benefit in the acute phases of crystalline-induced arthritis, especially gout. Since it is not ulcerogenic and has no effect on platelet function it would seem to have advantages in the treatment of the medical ICU patient. However, because the therapeutic dose is close to the toxic dose there is substantial risk of inducing vomiting or diarrhea in a patient population not well suited to withstand additional stress. While the intravenous preparation of colchicine is associated with less gastrointestinal toxicity, it is associated with more systemic toxicity and occasionally death. Risk factors for toxicity with intravenous colchicine include renal and hepatic insufficiency, advanced age, volume depletion, and underlying cytopenia. The use of intravenous colchicine should be avoided in critically ill patients.

Biological Agents

Tumor Necrosis Factor-α Inhibition

The selective targeting of the proinflammatory cytokine TNF-α has produced a therapeutic revolution in rheumatology. While most widely used in the management of the synovitis of rheumatoid arthritis, these agents have been widely applied to a variety of immune-mediated disorders including psoriatic arthritis, ankylosing spondylitis, Behçet's disease, and sarcoidosis. Three commercial products are available, including two monoclonal antibodies (infliximab [chimeric] and adalimumab [humanized]) and one recombinant TNF receptor Fc fusion protein (etanercept). There is an emerging consensus that the antibodies may be more effective in the non-RA patients, but perhaps not without additional toxicity. In general, these drugs are extremely well tolerated and are neither renal toxic or marrow suppressing. They can reactivate granulomatous infections, notably tuberculosis. They must be used cautiously in patients with chronic or recurrent infection. They should be discontinued in the setting of serious infection or sepsis. There is currently no established indication to begin TNF-directed therapy in critically ill patients. The observations that such therapy can benefit patients with refractory Wegener's granulomatosis and Behçet's disease with progressive visual loss suggest scenarios in which initiation of TNF-directed therapy might begin in the ICU.

Interleukin-1 Inhibition

The currently available therapy directed against the pivotal proinflammatory cytokine IL-1 is anakinra. It is a recombinant human IL-1 receptor antagonist. While moderately effective for rheumatoid arthritis, it has not been nearly as effective as the TNF-α drugs. Although it is associated with fewer serious infections than TNF-directed therapy, when they are used in combination the increased risk of serious infection outweighs any additional therapeutic benefit.

B-Cell Depletion

Rituximab is a chimeric (mouse/human) monoclonal antibody that targets the CD20 molecule on B cells. CD20 is unique to B cells and is a stable transmembrane marker expressed during all stages of their maturation. The monoclonal antibody depletes B cells by activating apoptosis signals and complement and Fc receptor cytotoxicity.87 Rituximab has an extensive history of effective use in B-cell lymphomas, and this bioengineered smart bomb has a surprisingly low number of significant side effects that can be characterized as constitutional symptoms of fever, chills, and rare hypotension attributable to cytokine release. All of the latter are most frequently experienced on the first infusion and are probably much more common in lymphoma patients with a large lymphocyte burden. The ability to selectively target and destroy B cells has generated tremendous interest in the rheumatologic community because rituximab offers a novel way to treat autoantibody-mediated autoimmune diseases like SLE, myasthenia gravis, autoimmune hemolytic anemia, and Goodpasture's syndrome.82 In critically ill patients who cannot tolerate or have failed alkylating agent therapy, rituximab offers the option of definitive suppression of autoantibody production. The creative use of rituximab in autoimmune disease will be limited only by the imagination of the rheumatologist, and the reality is that widespread off-label use is already in progress. However, it is doubtful that many of its diverse applications will be confirmed by appropriate clinical trials. In theory, the focused destruction of B cells should culminate eventually in hypogammaglobulinemia and put the patient at high risk for bacterial infection. However, this has not been the case in most patients with lymphoma who are treated with the standard protocol of two to six weekly infusions of 375 mg/m2. The treatment of patients with an autoimmune disease like SLE and the nephrotic syndrome may be a different situation since significant hypogammaglobulinemia requiring monthly replacement with IVIg has been reported. Careful monitoring of serum immunoglobulin levels at least in these patients is warranted over the 6- to 9-month period of rituximab-induced depletion.

Plasmapheresis

Therapeutic plasmapheresis can effect rapid removal of circulating antigens, immune complexes, pathologic antibodies, and circulating cytokines, and improvement in the function of the fixed monocyte system. Plasmapheresis should be considered for patients with fulminant vasculitis of any cause, drug-resistant thyroid storm, selected poisonings, and a subset of life-threatening pulmonary-renal syndromes. Because cyclophosphamide has at least a 7- to 14-day window prior to a therapeutic effect, plasma exchange can be efficacious as a temporizing measure in many syndromes that will ultimately require cytoreductive therapy. Plasma exchange is also the treatment of choice for patients with thrombotic thrombocytopenic purpura (TTP), acutely deteriorating myasthenia gravis, and rapidly progressive or severe inflammatory peripheral neuropathies. Therapeutic apheresis may also provide an immediate clinical response in some patients with fulminant SLE with severe CNS dysfunction. Cryoglobulinemia with deteriorating renal or nervous system disease and hyperviscosity syndrome are other disorders for which plasmapheresis may play an important therapeutic role.

The procedure has a wide margin of safety when contemporary machines are used. At any point in time the amount of extracorporeal volume is much less than when the patient is exchanged with equipment that requires centrifuge bowl separation of blood components. Still, patients with decreased effective plasma volume or significant coronary artery disease remain at risk with all plasma-separation technologies available and must be closely monitored during the procedure. The frequency of exchange can be based on the isotype of the putative pathogenic antibody involved and the known kinetics of IgG clearance. Ultimately, however, the frequency and amount of exchange depends on the clinical status of the patient. Plasma volume varies with the height, size, sex, and hematocrit; nomograms are available that should be used for effective calculation if the apheresis machine itself does not calculate the values automatically. A 1.0 plasma volume exchange will remove approximately 65% of intravascular IgG from a patient and 1.5 plasma volume exchange 78% to 80% of IgG. Larger volume removals are associated with increased time on the machine, increased citrate toxicity, and most importantly, only very small incremental increases in IgG removal. A common rationale used to determine the frequency of plasmapheresis is based on “resting” the patient on alternate days to allow re-equilibration between extravascular and intravascular IgG. TTP and hyperviscosity syndromes, in which plasma volumes are expanded and the proteins generating hyperviscosity are predominantly within the intravascular space, can be effectively treated by aggressive and daily higher volume exchanges.

Patients with severe hepatic disease cannot metabolize citrate rapidly, and plasma exchanges should be done at the lowest possible citrate:plasma ratio. Serious side effects of citrate-induced hypocalcemia are extremely rare when volume is replaced with calcium-spiked 5% albumin. Fresh frozen plasma should never be used for routine replacement unless the patient has TTP, concomitant severe clotting factor deficiencies, or will be undergoing major surgery within the next 24 to 48 hours. Temporary decreases in several clotting factors during plasma removal, reflected by changes in the PT and APTT, are usually transient and not of clinical concern unless the patient has significant associated hemostatic deficiencies or severe hepatic dysfunction.

Because many patients in ICUs will not have suitable antecubital venous access or will be unable to provide arm pumping to generate sufficient venous flow during an apheresis procedure, dual-lumen polyurethane dialysis apheresis catheters are usually placed in a central vein and will provide high-flow access. If subclavian or internal jugular catheters are used, the return port should be in the cava, since relatively undiluted citrated blood has on rare occasion been thought to precipitate arrhythmias. Most patients undergoing plasma exchange do not require intravenous antibody replacement after each apheresis procedure. Exceptions to this rule are patients with hypoproteinemic states, especially the nephrotic syndrome, those with active infection, and those undergoing intensive, daily, high-volume therapeutic aphereses.

A standard rule for apheresis treatment of patients with autoaggressive syndromes, especially those with autoimmune antibody production, is that plasma exchange alone, although an effective temporizing measure, will ultimately not be efficacious and theoretically could be deleterious to the patient unless the apheresis treatment is closely linked to the use of an alkylating agent. The theory behind coupled therapy is straightforward.82 Reduction in levels of pathogenic antibody, regulatory antibody, and regulatory cytokines can induce proliferation of B cells and possibly also T cells. Proliferating cells then become susceptible to alkylation, and either clonal abortion or anergy occurs when cyclophosphamide is given. The dose of cyclophosphamide ranges from 500 mg to 1 g/m2 in 5% dextrose and water over a period of 30 to 60 minutes. Premedication for nausea is given routinely but may be unnecessary in some patients. The alkylating treatment can be given immediately after the last plasmapheresis and up to 48 hours thereafter. Protocols for the frequency of apheresis, at least in the treatment of SLE, vary widely but usually encompass an induction period of five to seven exchanges of 1.0 to 1.5 plasma volume over 7 to 10 days until approximately 250 mL plasma per kilogram have been removed from the patient. This induction period is followed by some form of immunosuppressive therapy, usually cyclophosphamide.83

Protein a Columns

Protein A absorption columns provide on-line removal of IgG antibody and circulating immune complexes; they have the benefit of returning most other plasma components to the patient. Protein A pheresis has been sporadically effective in drug-resistant idiopathic thrombocytopenic purpura and cancer chemotherapy–associated TTP/hemolytic uremic syndrome.84 Angiotensin-converting enzyme inhibitors should be discontinued prior to the procedure. The efflux of activated inflammatory and vasoactive mediators generated from plasma by the protein A silica absorption vehicle can cause transient shaking, chills, fever, and hypotension. This usually occurs at the end or after the termination of the procedure.

Leukapheresis

Leukapheresis can be performed efficiently on centrifuge belt machines and is highly effective in reducing peripheral blast counts to levels that are less threatening to maintenance of adequate microcirculation and more amenable to chemotherapy. Combination leukaplasmapheresis has been performed in patients with scleroderma and polymyositis/dermatomyositis that was refractory to conventional management.

In puzzling cases the rheumatologist may discern from the nuances of the clinical examination, serologic testing, and invasive procedures that the patient has rheumatic disease. Sometimes the rheumatologist, like the intensivist, cannot make a definite diagnosis, yet confronts a critically ill patient who may have a rheumatic disease. The clinical status of such patients is usually at an unacceptable plateau or even more likely deteriorating. Should that patient be given empiric therapy? Recognition of our shortcomings in diagnosis and the inability of some patients to tolerate a critical invasive test recommends such a course in selected patients.

Prior to initiation of empirical therapy, it is helpful to ask a series of related questions:

1. Has infection been reasonably excluded? Infection and cancer most commonly mimic rheumatic disease. Since empiric therapy usually implies immunosuppressive drugs, most commonly corticosteroids, infection must be ruled out. Here the emphasis should be on reasonably making such an assessment. Endless sets of blood cultures should not delay a difficult decision.

2. What do I suspect the patient may have? It should be possible to formulate a plausible if unprovable diagnostic hypothesis. Such a hypothesis is critical to the rest of the experiment. If it is impossible to generate a “working diagnosis,” it is doubtful that the therapeutic trial will work. Autopsy study of these patients is likely to reveal cancer or no diagnosis.

3. What is adequate therapy for this suspected diagnosis? Treatment for a suspected diagnosis ranges from adequate to aggressive. In the absence of a definite diagnosis, it is reasonable to choose a level of therapeutic intensity that is usually adequate for the suspected disorder. Aggressive treatment approaches to an unconfirmed illness create another set of confounding variables and may place the patient at a further disadvantage.

4. What will I use as my parameters to judge therapeutic success? Empiric therapy should proceed with a clear understanding of the yardsticks that will measure therapeutic responsiveness. These parameters can then be rigorously monitored. Furthermore, blind spots in the baseline data can be addressed before therapy is begun.

5. What is the duration of a reasonable trial for this disorder? Agreement on the duration of a therapeutic trial should precede its initiation. Failure to develop such an end point can result in excessively long and risky therapy on the one hand, or a course that falls short of an adequate trial on the other. Furthermore, spontaneous improvement or improvement in response to other therapies may result in prolonged and unnecessary treatment.

The clinician needs to be vigilant about the risks associated with any trial of empiric therapy. However, after addressing the above questions, such a trial may represent both rational and compassionate care.