Dermatology for hospitalists: the OK, the bad and the really bad
Treating many cutaneous reactions is all about timing by Bonnie Darves
Published in the February 2011 issue of Today's Hospitalist
For hospitalists who can’t readily call a dermatology consult, figuring out what’s causing a rash in a patient who’s taking a slew of medications and has a host of comorbidities can be a challenge. But understanding the different types of cutaneous drug eruptions and identifying the drugs most likely to cause each type can help hospitalists cut to the chase, according to Lindy Fox, MD, assistant professor of clinical dermatology at University of California, San Francisco, and director of the dermatology hospital consultation service.
During a session on dermatology pearls at a UCSF
"Certain types of drug reactions occur only after certain periods of time."
–Lindy Fox, MD University of California, San Francisco
meeting for hospitalists held last year, Dr. Fox offered advice to help hospitalists deal expediently with drug-induced cutaneous eruptions and to differentiate between benign and serious reactions.
When timing is everything
According to Dr. Fox, cutaneous drug eruptions are really about statistics and timing. Hospitalists need to know both which high-risk medications are most often associated with a particular reaction and how long after a patient starts taking a drug an eruption tends to occur.
"Certain types of drug reactions occur only after certain periods of time," Dr. Fox said. "Even then, if there’s a high-risk medication on board, stop it even if you don’t think it’s causing the rash."
The short list of drug suspects in skin eruptions includes antibiotics, NSAIDs, sulfa drugs, allopurinol and anticonvulsants.
But what hospitalists need to remember, Dr. Fox stressed, is that the most common type of dermatological drug reaction— morbilliform, characterized by erythematous macules, papules and itching—typically doesn’t occur until seven to 10 days after a drug is started.
While hospitalists may be tempted to yank an antibiotic that a patient began taking one to two days before the cutaneous eruption began, the real culprit is often a medication that’s been on board longer. Hospitalists should suspect, for example, the Bactrim a patient started taking as an outpatient for a urinary tract infection just prior to admission.
The typical morbilliform reaction may look severe, but it has minimal systemic symptoms and no morbidity. The itchy rash typically occurs like a wave, affecting the trunk first, then the extremities. (It generally clears in that same order.)
Patients may look like they’re getting worse a few days into the eruption, Dr. Fox noted, but they are actually getting better. As the rash clears, patients may develop dry desquamation that looks like peeling skin after a sunburn.
"The skin underneath is actually normal," she said. "When there is no wet, oozy exposed dermis, there is nothing to worry about." A morbilliform eruption usually resolves within a week to 10 days after the drug has been stopped.
More serious reactions
Much more serious, however, are reactions that include drug-induced hypersensitivity syndrome (DIHS), Stevens-Johnson syndrome and toxic epidermal necrolysis (TEN). All can quickly become potentially life-threatening because they may cause organ damage. Although some studies put DIHS mortality rates as high as 25%, Dr. Fox thinks 5% to 10% is more accurate. TEN mortality is considerably higher, at 25% to 35%.
The heavy-hitter suspects in DIHS include anticonvulsants, sulfonamides, lamotrigine, NSAIDs, the HIV drugs abacavir and nevirapine, and minocycline.
One way to differentiate less serious eruptions from DIHS is to look at the drug-exposure time frame. DIHS typically occurs much later—two to seven weeks after a medication is started—than a typical morbilliform eruption.
With aromatic anticonvulsants like phenytoin, phenobarbital and carbamazepine, DIHS typically occurs about three weeks after exposure. With the sulfone dapsone, onset is roughly six weeks; for the hyperuricemia drug allopurinol, it’s seven weeks.
Signs and symptoms are also different. Early DIHS often looks like morbilliform, with a maculopapular eruption and, typically, itching skin. If the patient says her skin is painful, however, "something bad is happening," Dr. Fox said. "That is an immediate sign that skin is dying or an acute toxic reaction is occurring" and that Stevens-Johnson or TEN are more likely diagnoses.
In addition, patients may have facial edema, which doesn’t occur in morbilliform eruptions and is a warning that something serious is developing. Likewise, macules and papules begin to swell, weep and form tiny vesicles.
Early on, Dr. Fox observed, DIHS looks a lot like infectious mononucleosis. If patients on a high-risk drug develop fever, pharyngitis, lymphadenopathy, atypical lymphocytosis and liver-function abnormalities, hospitalists should suspect a hypersensitivity drug reaction. Eosinophilia is common and present in many but not all DIHS cases, so don’t consider it a requirement for a diagnosis.
Stevens-Johnson and TEN
When it comes to severe bullous drug eruptions (TEN and Stevens-Johnson syndrome), many consider them to be on a spectrum of the same disease. But Dr. Fox thinks they may actually be two different diseases.
Hospitalists should suspect Stevens-Johnson, which occurs in roughly six million people annually, when cutaneous detachment affects less than 10% of the body surface. Consider TEN, however, if the detached area is larger, 30% or more.
In addition, patients with Stevens-Johnson often report a prodrome, particularly fever, respiratory symptoms, headache, vomiting and diarrhea in the days before the eruption. TEN patients may have a burning sensation in their mucous membranes, particularly their eyes.
As a side note, Dr. Fox mentioned that Mycoplasma "is a really big player" in children and young adults with Stevens- Johnson, accounting for an estimated 25% of pediatric cases. She recommends always checking for Mycoplasma in that age group or in patients with severe mucous membrane involvement. Infrequently, herpes simplex virus may cause Stevens-Johnson, and the probability of Stevens- Johnson increases significantly in patients with HIV.
For severe bullous reactions, the same drugs that cause minor eruptions should be among the initial suspects. But another group of commonly prescribed medications recently linked to Stevens-Johnson and TEN include sertraline, pantoprazole and tramadol.
Patients with Stevens-Johnson typically develop targetoid lesions and have a relatively slower disease progression than those with TEN, Dr. Fox explained. "You have a little more time—a day or two—to watch," she said.
But some patients don’t develop targetoid lesions. They are instead "red from head to toe," report painful skin, and may start losing skin within a day, signs of a severe and acute form of TEN. The etiology of this type of TEN is almost always a medication and, as compared to Stevens- Johnson, lesions are dusky and poorly demarcated.
When the culprit is an aromatic anticonvulsant, hospitalists should stop the medication and avoid others in the same class. Safe alternatives include valproic acid or Keppra.
Treat DIHS with systemic steroids at a dose of 1.5 to 2 mg/kg daily and taper slowly, about 5% every three days, as patients improve.
"If you taper too quickly, they will rebound and be harder to control," she noted. The total taper may take six weeks or longer.
Allopurinol hypersensitivity may require other immunosuppressive therapy, such as Cellcept. Hospitalists dealing with an allopurinol reaction should avoid using azothioprine, which is also metabolized via the xanthine oxidase pathway.
If the DIHS diagnosis is made early enough, liver problems are generally completely reversible. But because some patients may develop hypo- or hyperthyroidism down the line, order a monthly thyroid-stimulating hormone test for six months. Dr. Fox also recommended following all lab results while patients are on steroids to make sure that the liver, kidney or bone marrow damage recovers.
Experts agree that all patients with suspected TEN or Stevens-Johnson should be treated emergently, Dr. Fox said. But precisely how to treat them is extremely controversial. While many practitioners insist that you should never use systemic steroids or IVIG for Stevens- Johnson or TEN, "I will tell you that you should," she said. "I use steroids for Stevens-Johnson and IVIG for TEN. But a lot will depend on where you practice and if you have a dermatologist who’s comfortable with these drugs."
For Stevens-Johnson patients, Dr. Fox recommended high-dose systemic steroids—1.5 to 2 mg/kg daily— and tapering slowly, about 5% every three days, as patients improve. Although the data are mixed, some studies have shown that IVIG treatment reduces TEN mortality. Dr. Fox recommended IVIG dosing of 0.5 to 1.0g/kg daily over three to five days. When the drug is effective, she said, it will arrest disease progression within 24 to 48 hours.
In addition, Dr. Fox advised the following management for suspected Stevens-Johnson or TEN:
Stop all unnecessary medications.
Arrange for an ophthalmology consult because of the high incidence of ocular morbidity.
Check for Mycoplasma, especially in pediatric patients.
Monitor fluid and electrolyte status and keep the environment warm.
Protect exposed skin by applying Aquaphor or Vaseline and covering with Vaseline-coated gauze.
Several experts also suggest avoiding the intensive rehydration used with burn patients because the depth of necrosis is less in TEN and Stevens-Johnson. "If you follow the burn formulas, you may end up over-hydrating them," Dr. Fox said. When patients are over-hydrated, they may become edematous, making it harder for damaged skin to heal.
Dr. Fox noted that there are situations with less severe reactions, such as morbilliform eruptions, in which an antibiotic cannot be safely stopped, as with treatment for endocarditis. When that occurs and the patient is near the end of the treatment, she leans toward continuing through week six.
"But if we need the drug lifelong, we’re in trouble, and will need to find an alternative," she said.
Bonnie Darves is a freelance health care writer based in Seattle.
A primer on purpura
WHEN IT COMES TO PURPURA—the spots and patches caused by bleeding into the skin that does not blanch with pressure—defining the kind of skin lesions patients have can help you decide what’s causing them.
Lindy Fox, MD, director of the dermatology hospital consultation service at University of California, San Francisco, gave the following overview of purpura at a UCSF meeting for hospitalists last year:
Nonpalpable purpura: This comes in two forms, depending on the size of the spots. Lesions as small as a pinpoint to 1 mm are petechiae; when they’re larger, about 3 mm or 4 mm, they’re considered to be macular purpura.
According to Dr. Fox, petechial eruptions are either platelet-related or not. "If they’re platelet-related, there is a problem with either platelet number or platelet function," she pointed out. Most non-platelet-related eruptions are caused by some kind of pressure, trauma or lack of support to the blood vessels. Those can include valsalva, trauma, scurvy, sun damage or steroid-induced atrophy, among others.
Whatever can give you petechiae can also cause macular purpura, so the differential diagnoses are very similar, Dr. Fox said. "But the lesions are bigger, so maybe something else is happening to the patient," she pointed out. "You now may have infection, inflammation or trauma superimposed on the other causes of petechiae." A common example in the hospital is patients with skin fragility or thrombocytopenia who get narrow red or purple streaks (vibex or linear purpura) from a blood pressure cuff.
Palpable purpura: Purpura that you can feel "usually implies inflammation around that vessel," Dr. Fox pointed out. "Classically, that is leukocytoclastic vasculitis." Causes for palpable purpura include idiopathic (45%- 55% of cases), infection (15%-20%), inflammatory diseases (15%-20%), medications (10%-15%) and malignancy (<5%).
Palpable purpura is often caused by immune complexes that include IgA vasculitis or Henoch-Schonlein purpura. In purpura not caused by immune complexes, the vessel damage is often mediated by a cell such as the neutrophil. Those tend to be ANCA-associated vasculitis, Sweet’s syndrome or leukemic vasculitis.
While you biopsy the skin to try to determine an infection, Dr. Fox said, "always do a blood culture because you may not find the infection on the skin."
Retiform purpura: These spots or patches have a netlike pattern, may or may not be palpable, and "are always due to vessel occlusion," said Dr. Fox. "The lesions can be small or large, depending on the etiology." In cases where patients have fever, erythema and only small amounts of purpura, "you probably also have infection or inflammation," she added.
With mostly purpura and little inflammation, on the other hand, "you probably have a primary occlusive process." That occlusion can be in the vessel wall as in vasculitis or in mineral deposition, such as calciphylaxis and oxalosis. Or it may be in the lumen, which could be caused by either a clot or an embolus.
"Cocaine exposure can also cause retiform purpura, either through a vascultis or a clogging of the vessel," said Dr. Fox. "A toxicology screen is now part of my work-up for somebody who has retiform purpura."