Tips from Other Journals
Pharmacologic Stress Testing in Women with Chest Pain
Noninvasive evaluation of women with chest pain and no previous history of coronary artery disease (CAD) is a challenge. Angina is about twice as likely to be the presenting symptom of CAD in women as it is in men but is associated with a high prevalence of normal coronary arteries in women. Exercise stress testing used alone as an assessment is unreliable because of the high incidence of false-positive tests in women. Very little information is available on the prognostic value of pharmacologic stress echocardiography in women with chest pain and no previous CAD. Cortigiani and associates studied the use of pharmacologic stress echocardiography in the evaluation of women who present with chest pain and have no previous history of CAD.
A total of 456 women who presented with a history of chest pain and no evidence of acute myocardial ischemia or injury were studied. None of the patients had a previous history of CAD, valvular disease, or dilated or hypertrophic cardiomyopathies. The patients underwent stress testing with either dipyridamole or dobutamine using standard protocols. The echocardiogram was interpreted by two experts, with a third expert's opinion included if there was significant disagreement about the interpretation. The patients were followed for about three years. Follow-up data included cardiac and noncardiac deaths, nonfatal myocardial infarctions, unstable angina and coronary revascularization.
No major complications occurred during the study. Fifty-one patients had an echocardiogram that was positive for ischemia, while 68 had a positive electrocardiogram with a negative echocardiogram. The remainder of the study subjects had negative results on both tests. A total of 23 cardiac events, three cardiac deaths, 10 acute myocardial infarctions and 10 unstable angina events occurred. No cardiac events occurred in the patients who had a positive electrocardiogram with a negative echocardiogram.
The authors conclude that the use of pharmacologic stress echocardiography in women with chest pain and no previous history of CAD can differentiate between low or high risk in women who are at risk for cardiac events. In a patient with a negative test, the chance of having a cardiac event over a three-year period is less than 1 percent. Women with an abnormal electrocardiogram but a normal echocardiogram can be classified as low risk.
KARL MILLER, M.D.
Cortigiani L, et al. Prognostic value of pharmacological stress echocardiography in women with chest pain and unknown coronary artery disease. J Am Coll Cardiol December 1998;32:1975-81.
Explaining Gout to Patients by Using an Analogy
Treatment of Gout
Acute attacks: "Agents that put out the fire"
Indomethacin, 50 mg three times a day until all symptoms have resolved for 48 hours (any other NSAID may be substituted and used at the maximum approved dosage); or colchicine, 0.5 or 0.6 mg every hour until relief or intolerable side effects (abdominable cramping or diarrhea) occur, or until 10 doses have been taken without improvement; or corticosteroid intra-articular injections; or ACTH, 40 U intramuscularly, up to three times a day; or triamcinolone acetonide, 60 mg intramuscularly; or prednisone, 30 to 50 mg daily, tapering over seven days.
Prophylaxis against acute attacks: "Make the matches damp"
Colchicine, 0.5 or 0.6 mg, one to three times a day. Generally, this agent should not be employed unless the patient is also taking a urate-lowering drug; colchicine is usually discontinued after the serum urate concentration has been controlled and the patient has not had an acute attack for one to three months.
Urate-lowering agents: "Remove the matches from the body"
Allopurinol, generally 300 mg daily.
Uricosuric agents; probenecid, 500 to 1,000 mg twice daily, or sulfinpyrazone, 50 to 400 mg twice daily or benzbromarone, 25 to 120 mg daily.
The dosages of any urate-lowering agent are the lowest needed to maintain the serum urate concentration at or below 5 mg per dL.
NSAID=nonsteroidal anti-inflammatory drug; ACTH=adrenocorticotropic hormone.
Adapted with permission from Wortmann RL. Effective management of gout: an analogy. Am J Med 1998;105:513-4.
Gouty arthritis is not a difficult diagnosis to make. Patients often suffer from recurrent attacks because of the complexity of the treatment regimen, with the potential for three medications to be taken on different schedules. Wortmann discusses an analogy that may make understanding gout easier for both patients and physicians.
The analogy used is, "Gout is like matches." Everyone has uric acid in their blood; high levels of uric acid may cause gout. When too much uric acid accumulates around a person's joints, the uric acid acts like matches. An acute gout attack may be likened to "setting the joint on fire." A nonsteroidal anti-inflammatory drug (NSAID) should be taken at this point, to prevent more matches from catching on fire. It isn't important which NSAID is used to "put out the fire," but how quickly the fire is put out is very important. The more matches that catch fire, the longer it will take to put the fire out. However, if the matches still exist, they can catch on fire again--so NSAIDs should not be seen as a "cure" for gout. A urate-lowering medication, such as allopurinol, is used to "remove the matches." In the analogy, no matches means no fire, which means no gout attacks. However, before all the matches have been taken away, they can still catch on fire. What is needed is something to "dampen the matches," so that they cannot catch fire as easily. Colchicine is an agent used in this way to help prevent acute attacks.
Compliance often improves when the patient has a better understanding of the disease process; the analogy presented here may help patients understand and accept the steps they need to take in the treatment of gout. The accompanying table gives the standard treatment of gout, along with the corresponding phrase in the "fire" analogy.
GRACE BROOKE HUFFMAN, M.D.
Wortmann RL. Effective management of gout: an analogy. Am J Med December 1998;105:513-4.
Which Antibiotic to Choose for Recurrent Otitis Media?
Controversy about appropriate antibiotics, length of treatment and whether to treat acute otitis media continues. In addition, when antibiotic therapy has been chosen as the most appropriate course and a primary treatment failure occurs, the choice for repeat treatment may not be clear. Hueston and colleagues reviewed the antibiotics being used in these cases and evaluated reports to see which medications were most likely to be effective and result in fewer treatment failures.
The medical records of children younger than 12 years with otitis media were reviewed. The patients who returned within 45 days of the initial diagnosis for otitis media and received a new prescription for an antibiotic were classified as primary treatment failures. Those with primary treatment failure were reviewed to determine which of these children subsequently had a 90-day period without otitis media, followed by a new diagnosis of otitis media. The antibiotic used in the final episode of otitis media was recorded, as was the medication chosen for the first episode. First-line drugs were amoxicillin, ampicillin, penicillin and trimethoprim-sulfamethoxazole. All other drugs were considered second-line medications.
A total of 7,807 children were diagnosed with otitis media. Of these, 1,416 patients (18 percent) were classified as having a primary treatment failure. Subsequently, 343 patients were seen for recurrent otitis media occurring at least 90 days after the primary treatment failure. A first-line medication was prescribed in 69 percent of these cases. Second-line antibiotics were more likely to be used if the recurrence occurred soon after the treatment failure. Rates of failure were similar for first- and second-line antibiotics (13 percent vs. 19 percent, respectively).
The authors conclude that there is no benefit to choosing a second-line antibiotic (a broad-spectrum antibiotic) when otitis media recurs following a treatment failure. When a child presents with a recurrence of otitis media after a presumed treatment failure, amoxicillin, ampicillin or trimethoprim-sulfamethoxazole are reasonable treatment choices that will be less costly, just as effective and less likely to cause resistant strains of micro-organisms than would use of broad-spectrum agents.
GRACE BROOKE HUFFMAN, M.D.
Hueston WJ, et al. Treatment of recurrent otitis media after a previous treatment failure. Which antibiotics work best? J Fam Pract January 1999;48:43-6.
Horse-Chestnut Seed Extract for Chronic Venous Insufficiency
Therapeutic response to mechanical compression for chronic venous insufficiency is often less than expected because of poor compliance. For this reason, an oral therapy would be desirable. Horse-chestnut seed extract has been studied mainly in Europe as a possible oral therapy for chronic venous insufficiency. It may work by preventing leukocyte activation, which is considered a pathophysiologic mechanism of chronic venous insufficiency. Pittler and Ernst conducted a literature review to evaluate the evidence for or against the use of this compound for venous insufficiency.
The authors found 13 randomized controlled trials of horse-chestnut seed extract: eight were placebo-controlled trials and five were controlled against reference medications. The number of subjects in the placebo-controlled trials ranged from 20 to 226, and the studies lasted from 20 days to eight weeks. In the five trials controlled against reference medications, the size of the studies ranged from 30 to 240 subjects, and the studies lasted from four to 12 weeks.
The findings of the placebo-controlled trials suggested that treatment with horse-chestnut seed extract produced a decrease in lower-leg volume and calf and ankle circumference. Edema provocation before and after the treatment period revealed protective effects against edema. One study demonstrated a 22 percent decrease in the capillary filtration rate in patients who received horse-chestnut seed extract. Similarly, all five of the randomized control trials against reference medications demonstrated evidence for the effectiveness of horse-chestnut seed extract in the treatment of chronic venous insufficiency. One trial that compared the herbal medication to compression stockings suggested a therapeutic equivalence, but the authors noted that this trial was not properly blinded, so bias may have affected the results.
Data from the studies showed that therapeutic benefit was derived from doses of horse-chestnut seed extract standardized to 100 to 150 mg of escin daily. (Escin is the active component of the extract.) In two of the placebo-controlled studies, daily doses of 100 mg of escin resulted in a significant reduction in the mean leg volume after two weeks of therapy. In another study, the effects of treatment were noted to persist: at the end of a six-week follow-up period the mean leg volume was not significantly different from that immediately after treatment.
The authors note that many experts remain skeptical about the effectiveness of horse-chestnut seed extract for chronic venous insufficiency. Most of the studies the authors analyzed were from German investigators. Further trials are needed. If further studies substantiate the benefits of horse-chestnut seed extract, its postulated mechanism of action might generate interest in other forms of drug therapy. Escin has been shown in vitro to inhibit the activity of elastase and hyaluronidase. Both of these substances are involved in enzymatic proteoglycan degradation, which is part of the capillary endothelium and the main component of the extravascular matrix. Horse-chestnut seed extract may prevent vascular leakage by shifting the equilibrium between the degradation and synthesis of proteoglycans.
BARBARA APGAR, M.D., M.S.
Pittler MH, Ernst E. Horse-chestnut seed extract for chronic venous insufficiency: a criteria-based systematic review. Arch Dermatol November 1998;134:1356-60.
Use of Tissue Adhesive Versus Suture Wound Repair
Indications for Use of Octylcyanoacrylate in Wound Closure
Anatomic location
Indication
Face Most cutaneous closures Lips and mucosa Not recommended Extremities and torso Cutaneous closures, deep sutures recommended; not over joints Hands and feet Minor lacerations only; generally not recommended
Reprinted with permission from Quinn J, Wells G, Sutcliffe T, Jarmuske M, Maw J, Stiell I, et al. Tissue adhesive versus suture wound repair at 1 year: randomized clinical trial correlating early, 3-month, and 1-year costmetic outcome. Ann Emerg Med 1998;32:645-9.
Traumatic wounds have traditionally been closed with sutures. Recent trials comparing the use of sutures with tissue adhesives have demonstrated similar cosmetic outcomes. To date, all previous studies have compared outcomes between sutures and tissue adhesives at three months after wound closure. Quinn and associates compared traumatic lacerations treated with octylcyanoacrylate, a tissue adhesive recently labeled for this use by the U.S. Food and Drug Administration, with lacerations repaired with sutures.
The one-year study included 135 adult patients with traumatic lacerations requiring repair on the face, torso or extremities. Contaminated wounds, scalp or ear lacerations, and extremity lacerations in patients with diabetes or chronic steroid use constituted exclusion criteria. Wounds requiring debridement or deep suturing were first treated in accordance with the standard practice of the treating physician. Only at the time of skin closure were the lacerations randomized to either closure with monofilament or octylcyanoacrylate tissue adhesive. Patients with wounds requiring no deep care were immediately assigned to one of the two closure techniques. All study physicians were trained in the use of tissue adhesives by the study's lead investigator.
Wound dehiscence was treated by primary closure with the previously randomized method of closure. Follow-up was conducted within 10 days, then at three months and one year after wound repair. All wounds were evaluated at follow-up visits using a previously validated wound-evaluation scale. Photographs taken by a research assistant at three months and one year were rated for cosmesis by a cosmetic surgeon. Of the 77 patients who were available at the three-month and the one-year follow-up visits, there was no difference in primary outcome, visual analog cosmesis score or percentage of optimal wounds score between the two treatment groups.
The authors conclude that there is no difference one year after treatment in the cosmetic outcomes of wounds repaired with tissue adhesives and those closed with sutures. Their findings confirmed earlier similarities in cosmetic results noted after three months. Physicians who use octylcyanoacrylate should know its proper indications and method of application (see the accompanying table). Topical anesthetics help to achieve hemostasis, permit proper cleansing and allow a needle-less closure. Tissue adhesives should always be applied topically, never between the wound edges.
RICHARD SADOVSKY, M.D.
Quinn J, et al. Tissue adhesive versus suture wound repair at 1 year: randomized clinical trial correlating early, 3-month, and 1-year cosmetic outcome. Ann Emerg Med December 1998;32:645-9.
EDITOR'S NOTE: Cyanoacrylate tissue adhesives have demonstrated satisfactory wound healing and cosmetic results, while requiring less time to initially close a wound than suturing. Subjective parental reporting of perceived pain among children undergoing wound closure is significantly less with the chemical tissue adhesive. Application of the adhesive can be difficult because the low viscosity of the formulas causes "running" of the adhesive on initial application. This "running" can be diminished by using a needle placed in the top of the tube or by releasing the adhesive down a needle to provide more exact application. Future formulas will be more "user friendly."
R.S.
Antidepressants Are Effective in Treatment of Adult ADHD
Diagnosis of attention-deficit/hyperactivity disorder (ADHD) in adults is difficult. Some evidence indicates that most children outgrow this disorder by adulthood, but other evidence shows that some adults may continue to have at least residual symptoms of ADHD. Higgins reviewed the literature to assess the efficacy of antidepressants compared with stimulants in adults who presented with attention deficits.
A Medline search yielded 20 studies that reported on antidepressant or stimulant treatment in adults with attention deficits. These studies were reviewed to determine the type of medication used and the percentage of patients who were responders. Of the nine studies of stimulants, only five were placebo-controlled double-blind studies. Three of these reported significant improvement in patients taking the stimulant compared with placebo; the other two studies showed no difference between the two.
Only three of the 11 studies of antidepressants were placebo-controlled. One study evaluated the use of a monoamine oxidase inhibitor and showed no difference between the active treatment and placebo. The other two studies (using desipramine and tomoxetine, a medication that is not available in the United States) found that 68 percent and 52 percent of patients responded, respectively, compared with none and 2 percent of those in the placebo group. Nonplacebo-controlled studies of bupropion and venlafaxine found good response in adults with attention deficits.
Further studies, including those comparing antidepressants, stimulants and placebo, are needed to determine the best treatment in adults with ADHD. However, the author notes that using an antidepressant in adequate dosages may be a satisfactory alternative to prescribing a controlled substance (i.e., a stimulant) in these patients.
GRACE BROOKE HUFFMAN, M.D.
Higgins ES. A comparative analysis of antidepressants and stimulants for the treatment of adults with attention-deficit hyperactivity disorder. J Fam Pract January 1999; 48:15-20.
Behavioral Treatment Effective for Urge Incontinence
Urge urinary incontinence is most often treated with pharmacologic therapy, with oxybutynin frequently prescribed. Burgio and colleagues conducted this randomized trial to compare oxybutynin treatment with behavioral biofeedback-assisted treatment in women with urge incontinence.
Women who were at least 55 years of age and had urge incontinence at least twice a week for at least three months were eligible. Patients were excluded if they were cognitively impaired, if their postvoid residual volume was greater than 200 mL or if they had medical conditions that would preclude the use of oxybutynin, such as glaucoma, heart problems or a prolapsed uterus. Patients were classified as having mild, moderate or severe incontinence (based on the number of incontinent episodes per week) and were then randomized to one of three groups: the control group, the behavioral treatment group and the medication group. Patients in the behavioral treatment group received anorectal biofeedback training, learned strategies for reducing the urge sensation and were instructed in home practice of pelvic muscle exercises. The medication group received oxybutynin in a dosage of 2.5 mg, three times a day, with dosage adjustment to minimize side effects while maximizing effectiveness.
All patients completed four clinic visits at two-week intervals for an eight-week period. At each visit, patients submitted a diary of daily incontinence episodes, completed an adverse event questionnaire and submitted a urine specimen. Vital signs were also recorded, and anal sphincter pressure was measured. Patients were asked after the study to rate their own progress and satisfaction.
The study randomized 197 women, with 65 each in the behavioral and control groups and 67 in the medication arm. Not unexpectedly, the medication group reported a higher incidence of side effects (almost 97 percent reported dry mouth). The control group had the highest rate of incontinence at the end of the study, with the behavioral group reporting the lowest rate of incontinence. From baseline to posttreatment, there was an 80.7 percent reduction in incontinence episodes in the behavioral group, compared with a 68.5 percent reduction in the medication group and a 39.4 percent reduction in the control group. Patients in the behavioral group were the most satisfied with their treatment; only 14 percent wished to add medication treatment to their behavioral regimen. In contrast, 75.5 percent of those receiving oxybutynin wished to receive another form of treatment.
The authors conclude that behavioral intervention for the treatment of urge incontinence in older women is safe, effective and well-liked by patients. Of note, the control group also achieved a significant reduction in incontinence episodes, showing that attention from the physician and expectations about the possibility of improvement can, in and of themselves, be beneficial.
In an accompanying editorial, Resnick points out that individualized attention (be it behavioral or pharmacologic) may be the most important factor for women with urge incontinence. However, he does note that the Burgio study was restricted to a specific cohort in which the women were cognitively intact, healthy and fairly young. In addition, some of the patients included had some element of stress incontinence, which may lend itself somewhat more to a behavioral approach.
GRACE BROOKE HUFFMAN, M.D.
Burgio KL, et al. Behavioral vs. drug treatment for urge urinary incontinence in older women. A randomized controlled trial. JAMA December 16, 1998;280:1995-2000, and Resnick NM. Improving treatment of urinary incontinence [Editorial]. JAMA December 16, 1998;280:2034-5.
Use of Repaglinide Along with Metformin for Glycemic Control
Repaglinide augments insulin release when it is taken with meals, an effect that may serve to complement the effects of drugs that decrease insulin resistance. Repaglinide has a short half-life and is excreted in the bile, making it particularly suitable for patients with renal impairment and for elderly patients. Moses and colleagues studied the efficacy of repaglinide as a substitute agent in patients who cannot take metformin or as a second agent in patients whose plasma glucose levels are not optimally controlled with metformin alone.
The double-blind, placebo-controlled study included 83 patients with type 2 diabetes (formerly known as non-insulin-dependent diabetes) and hemoglobin A1c (HbA1c) levels of at least 7.1 percent (mean HBA1c: 8.5 percent) despite at least six months of metformin therapy. Patients were randomized to one of three treatment groups: continuation of their current dosage of metformin (27 patients); a change to repaglinide monotherapy (29 patients); or the addition of repaglinide to their current dosage of metformin (27 patients). Seventy-four patients completed the trial: 21 in the metformin group, 26 in the repaglinide group and 27 in the combination therapy group.
The optimal dosage of repaglinide was determined by titrating the dosage upward every four to seven days until the patient had symptoms of hypoglycemia or a fasting plasma glucose of less than 140.5 mg per dL (7.8 mmol per L). The repaglinide dose was decreased if the plasma glucose level dropped to below 79.2 mg per dL (4.4 mmol per L) or if clinically significant hypoglycemia developed. Patients continued their drug regimen for three months after a four- to eight-week period of titration.
No change in the HbA1c level or fasting plasma glucose concentration occurred in patients who continued receiving metformin or changed to repaglinide. However, both the fasting plasma glucose and HbA1c levels decreased significantly in patients who took repaglinide in addition to metformin. Combination therapy was accompanied by a decrease in the HbA1c concentration from 8.3 percent to 6.9 percent. The fasting plasma glucose decreased from 183.7 mg per dL (10.2 mmol per L) to 144 mg per dL (8.0 mmol per L). Furthermore, none of the patients entered the study with optimal control of diabetes; 25 percent had HbA1c levels of greater than 9 percent. However, by the end of the study, nearly 60 percent of the subjects who received combination therapy were in optimal glycemic control (HbA1c of less than 7.1 percent). In contrast, optimal control was achieved in only 20 percent of those receiving repaglinide or metformin alone. None of the patients in the combination therapy group had an HbA1c level above 9 percent at the end of the trial.
There were 78 mild or moderate adverse events that were considered to possibly or probably relate to repaglinide or metformin. The most common adverse effects were hypoglycemia, diarrhea and headache. Of the 39 episodes of symptomatic hypoglycemia reported in the repaglinide and combination therapy groups, 44 percent (i.e., 17 of the 39 episodes) occurred during titration. Twelve of the 17 hypoglycemic episodes occurred in a single patient.
The authors conclude that the addition of repaglinide improves glycemic control in patients who have not achieved sufficient control with metformin alone. In addition, the findings show that repaglinide can be substituted for metformin without a change in glycemic control. The authors point out that combination therapy with metformin and repaglinide may be useful in patients who experience unacceptable side effects from metformin. In such a situation, repaglinide could be added and the metformin dose could be reduced without jeopardizing glycemic control. Although there were more episodes of hypoglycemia in the combination therapy group compared with the single-drug groups, such an increase in hypoglycemic episodes would be expected to occur with tighter glycemic control. The short half-life of repaglinide may explain why the mild to moderate episodes of hypoglycemia did not progress to severe episodes from prolonged hyperinsulinemia.
CLARISSA C. KRIPKE, M.D.
Moses R, et al. Effect of repaglinide addition to metformin monotherapy on glycemic control in patients with type 2 diabetes. Diabetes Care January 1999;22:119-24.
Novelty Enhances Success in Many Weight Loss Efforts
Many of the factors that lead to a successful weight loss diet remain unclear. In controlled trials, patients in the placebo group often achieve impressive results. These results may be due to the ambiance of the trial or weight loss program, particularly regular contact with staff, and the belief that the treatment is novel and likely to be successful. Summerbell and colleagues studied the role of novelty and simplicity in weight loss programs.
Forty-five adults referred to an obesity clinic were included in the study. Patients had a body mass index of more than 27, were willing to follow diet management and had no contraindications to calorie restriction. Patients were excluded from participation if they had diabetes or were pregnant or lactating. The patients were randomly assigned to follow one of three diets for 16 weeks. Each of the diets was designed to produce an energy deficit of four to seven mega-joules per day. The patients in the control group were instructed in a conventional balanced diet of normal foods. Patients following the novel "milk-only" program were restricted to a simple diet of milk and unsweetened yogurt. The third group of patients followed a "milk-plus" diet in which they could select a favorite food to supplement a basic milk diet.
Nine of the 14 control subjects completed the 16-week program. Eleven of 14 study subjects completed the milk-only diet and 11 of 17 subjects completed the milk-plus diet. The mean weight loss was 2.6 kg (5.7 lb) in the control subjects, 11.2 kg (24.6 lb) in the milk-only group and 8.2 kg (18.0 lb) in the milk-plus group. The results for the milk-only diet were superior to reported losses in trials of appetite suppressant and other dietary medications.
The authors emphasize that the trial took place under realistic conditions. Patients had no incentive except their own wish to lose weight. They bought supplies at their usual grocery stores, used no medical or surgical adjuvant treatment and were not provided with additional exercise or behavior therapy. No side effects were reported except constipation, and no evidence of nutritional deficiency was noted. The authors attribute the success of the restricted diets to novelty and simplicity. Compliance was higher for the two novel diets than for the conventional diet.
The authors conclude that novelty and simplicity are key to the success of a short-term diet for weight reduction. They do not recommend these diets for prolonged periods but emphasize that these novel diets have results that rival those of drug treatments. The authors suggest that various diets could be rotated to sustain patient confidence during prolonged weight loss.
ANNE D. WALLING, M.D.
Summerbell CD, et al. Randomised controlled trial of novel, simple, and well supervised weight reducing diets in outpatients. BMJ November 28, 1998;317:1487-9.
Can Tube Feeding Cause Diarrhea from C. difficile?
Clostridium difficile is the most common cause of nosocomial diarrhea. Hospitalized patients with tube feeding often have diarrhea, but C. difficile has not been well studied as an etiologic agent of these cases of nosocomial diarrhea. To examine this, Bliss and associates used cultures and cytotoxin assays to prospectively study the incidence of C. difficileassociated diarrhea in tube-fed patients and in a matched set of nontube-fed patients in an acute care setting.
Patients with previous or baseline positive C. difficile cultures were excluded from the study. Stool frequency and consistency were recorded, and rectal swab cultures were monitored in 76 tube-fed patients and 76 matched nontube-fed patients who served as control subjects. When diarrhea occurred, stool cultures and cytotoxin assays were obtained every 48 hours until a positive test result was obtained or until the diarrhea ceased.
Overall acquisition of C. difficile was significantly greater in tube-fed patients (15 of 76 patients, or 20 percent) than in nontube-fed patients (six of 76 patients, or 8 percent). C. difficileassociated diarrhea occurred in seven (9 percent) tube-fed patients and one (1 percent) matched control patient. The total incidence of diarrhea from all causes was higher in the tube-fed group (38 of 76 patients, or 50 percent) than in the control group (16 of 76 patients, or 21 percent). The incidence of C. difficileassociated diarrhea was higher in tube-fed patients who received postpyloric tube feedings during the study than in those who received prepyloric tube feeding continuously. Exposure to antibiotics was similar in patients who acquired C. difficile and in those who did not, but more patients with C. difficileassociated diarrhea received third-generation cephalosporins and aminoglycosides than patients without this diarrhea.
The authors conclude that tube feeding is a risk factor for acquiring C. difficile and developing C. difficileassociated diarrhea. This difference could not be attributed to antibiotic use since so many patients received antibiotics. It may be caused by transmission of bacteria from the hands of health care providers during routine manipulation of the tube-feeding system, from contaminated formulas and delivery systems, or from a lack of dietary fiber. Postpyloric tube feedings are associated with a greater incidence of C. difficile acquisition and diarrhea because feeding below the gastric acid barrier may facilitate organism survival. Tube-fed patients with diarrhea should be routinely tested with stool cultures for C. difficile.
RICHARD SADOVSKY, M.D.
Bliss DZ, et al. Acquisition of Clostridium difficile and Clostridium difficileassociated diarrhea in hospitalized patients receiving tube feeding. Ann Intern Med December 15, 1998;129:1012-9.
Does Weight Loss Plus Exercise Improve Insulin Sensitivity?
Weight loss in obese persons is associated with a decrease in insulin resistance and postprandial glucose and insulin levels. It has been suggested that additional benefit may be derived by combining weight loss and exercise. Weinstock and associates investigated the effects of diet and exercise on weight loss and insulin sensitivity in 45 obese women without diabetes.
The subjects were randomly assigned to one of three 48-week weight loss programs: diet alone; diet and aerobic training; and diet and strength training. All women participated in the same group behavior modification program and diet program. They consumed a diet of 925 calories per day for the first 16 weeks, followed by an increase to 1,500 calories per day for the remainder of the supervised diet and exercise program. Exercise consisted of three sessions per week for the first 28 weeks and two sessions per week for the next 20 weeks. Exercise was unsupervised during the remainder of the follow-up period.
Twenty-two subjects were also evaluated approximately one year after the study (week 96). To assess the effects of weight loss and exercise on insulin sensitivity, oral glucose tolerance tests were performed at baseline and at weeks 16, 24, 44 and 96.
Subjects in all three groups lost weight during the first 16 weeks. At week 16, the mean weight loss was 13.8 kg (30.4 lb), and this weight loss was maintained through weeks 24 and 44. In the 22 subjects who returned for a final visit at week 96, weight had increased from the 44th week to the 96th week, resulting in a mean net weight loss of 9.9 kg (21.8 lb) from baseline weight. At week 44, these subjects demonstrated a mean 15.2-kg (33.4-lb) weight loss. From weeks 44 to 96, during the unsupervised period, 14 of the 22 subjects (64 percent) gained more than 5 kg (11 lb). No significant differences were observed among the women in the three diet and exercise groups at week 96.
Assessment of glucose tolerance during the study period revealed that fasting glucose levels and glucose levels obtained after a 75-g glucose load did not differ among the groups throughout the study. The mean fasting insulin level and the mean insulin level in response to oral glucose decreased significantly from baseline to the 44th week, after weight loss had been achieved. The type of exercise program (i.e., aerobic or strength training) or the lack of an exercise program did not have a bearing on insulin levels.
In the 22 subjects followed for 96 weeks, fasting insulin levels and insulin levels after the glucose load rose from baseline in 19 of the subjects (86 percent) so that at week 96 their insulin levels were not significantly different from baseline levels. This was in contrast to weight, which remained significantly lower than baseline weight.
The authors conclude that their study corroborates the benefit of weight loss on hyperinsulinemia in obese persons. The addition of exercise, however, was not found to provide additional improvement, and a marked increase in insulin levels was noted with only a partial regaining of weight. Further studies are needed to investigate whether insulin sensitivity can be improved with more sustained weight loss or whether a defined basal metabolic index is required to improve insulin sensitivity.
BARBARA APGAR, M.D., M.S.
Weinstock RS, et al. Diet and exercise in the treatment of obesity: effects of 3 interventions on insulin resistance. Arch Intern Med December 1998;158:2477-83.
Varicella Vaccine: Alone or With MMR and DTP/HbOC?
The recent increase in serious disease caused by varicella and resulting complications, in both children and adults, have made its prevention desirable. Varicella vaccine, which was licensed in March 1995, is recommended by the Advisory Committee on Immunization Practices and the American Academy of Pediatrics. Although administering multiple vaccines at a single visit is convenient, lower varicella antibody response has been seen when varicella vaccine was combined with measles-mumps-rubella (MMR) vaccine and administered in a single vaccine as MMRV. Shinefield and colleagues studied the immunogenicity and safety of varicella vaccine administered concurrently at a separate site with MMR and a combination diphtheria, tetanus toxoids and pertussis (DTP)Haemophilus influenzae type b vaccine (DTP-Hib conjugate combination, or DTP/HbOC) as opposed to single administration of varicella vaccine six weeks after concomitant administration of MMR and DTP/HbOC.
A total of 609 infants were randomly assigned to one of two groups: one receiving MMR, DTP/HbOC and varicella vaccines at the same time but at separate injection sites, and the second receiving varicella vaccine six weeks after the concomitant administration of MMR and DTP/HbOC vaccines. All participants were followed clinically for local or systemic reactions to injection. Serologic immunities were measured approximately one year after vaccination.
Significantly more pain and soreness occurred when varicella vaccine was given concomitantly than when it was given six weeks after the other two vaccines. Irritability and temperature of at least 38.°C (102°F) was significantly more frequent following concomitant administration of the three vaccines. At one-year follow-up, mean titers were similar in both groups for all viral antigens. Six cases (three from each group) of varicella occurred within 42 days of vaccination. Two cases (both from the first group) of varicella were reported after 42 days of vaccination.
The authors conclude that since the antibody responses and the number and severity of varicella cases occurring within 42 days of vaccination were similar following both courses of immunization, the administration of varicella vaccine, MMR and DTP/HbOC at the same visit but at separate injection sites is as effective as varicella vaccine administration at a separate visit.
RICHARD SADOVSKY, M.D.
Shinefield HR, et al. Safety, tolerability and immunogenicity of concomitant injections in separate locations of M-M-RII, Varivax and Tetramune in healthy children vs. concomitant injections of M-M-RII and Tetramune followed six weeks later by Varivax. Pediatr Infect Dis J November 1998;11;980-5.
EDITOR'S NOTE: Administration of varicella vaccine has been hindered by both physicians' and parents' concerns about efficacy, length of immunity and side effects, and a perception that chicken pox is a benign disease. Studies in the United States show good humoral and cellular immunity up to 10 years after vaccination, while experience in Japan shows maintenance of active immunity by vaccination for 20 years in over 90 percent of initial responders. Safety has been documented, with the most common adverse effect being a brief, localized reaction to the injection. A few cases of mild symptoms with a varicella-like rash have been reported. Physicians need to more strongly advocate for vaccination of susceptible children and adults, since chicken pox is not always benign, especially if contracted after childhood. The World Health Organization has noted that universal administration of this varicella vaccine may not be a high priority for developing nations where cost considerations are more important.
R.S.
New Developments in the Management of Chronic Pain
Chronic pain is sustained irrespective of the original cause, and blocking or severing neural pathways merely shifts pain stimuli more centrally. Nurmikko and colleagues review the management of chronic pain, including developments in our understanding of the pathophysiology and pharmacology of chronic pain and selection management strategies.
Chronic neuropathic pain is believed to result from altered neuronal excitability resulting in abnormal discharges and central sensitization. Anticonvulsant drugs such as lamotrigine and gabapentin may be useful in the management of this type of pain. Studies validate the effectiveness of lamotrigine in allodynia and hyperalgesia, as well as trigeminal neuralgia (when used in combination with carbamazepine). Central neuropathic pain may respond to high dosages of lamotrigine (up to 600 mg per day), but the risk of side effects increases at high dosages. Gabapentin has been effective in studies of diabetic neuropathy and postherpetic neuralgia and is favored by many clinicians because it has few interactions with other drugs. The most common side effect is sedation.
The most frequently used external agents for chronic pain are capsaicin and transdermal fentanyl. Concentrations of capsaicin in ointment vary from 0.025 to 0.075 percent. The ointment causes local erythema but is reported to be effective in cases of diabetic neuropathy, postherpetic pain and other types of neuropathic pain. Transdermal fentanyl patches are suitable for patients with cancer pain who have difficulty swallowing. The patches release 25 to 100 µg per hour and are applied every 72 hours. Although their effectiveness is well-established in cancer therapy, the use of these medications for other types of chronic pain remains controversial. Side effects include constipation and hypoventilation.
Controlled studies of spinal cord stimulation are difficult to conduct. The literature suggests that spinal cord stimulation will relieve pain in approximately one half of cases of peripheral neuropathic pain and back pain persisting following surgery, in 80 percent of patients with intractable anginal pain, in 50 to 80 percent of patients with peripheral limb ischemia and in selected patients with diabetic neuropathy. The spinal cord stimulation procedure is expensive, with implant costs exceeding $10,000, and further expenses for prolonged monitoring., However, there is some evidence that the expense is cost effective in the long term.
Several analgesic agents are in development or undergoing clinical trial. Tramadol, a centrally acting opiate, has a low potential for addiction and does not interact with anticoagulants or irritate gastric mucosa. Selective and preferential inhibitors of cyclo-oxygenase 2 are also being developed to provide analgesia without gastrointestinal damage.
Despite use of optimal regimens and combinations of analgesics and pain-relieving techniques, many patients continue to have psychologic and physical disability because of chronic pain. Comprehensive pain management programs that incorporate relaxation therapy, education, cognitive therapy and a variety of other techniques have generated varied results, as measured by patients' ability to return to normal activities.
The authors conclude that the management of chronic pain continues to be challenging despite many promising developments. Changes in attitudes and in the organization and accessibility of health care services, as well as advances in therapy, will be needed to provide effective relief for patients with chronic pain.
ANNE D. WALLING, M.D.
Nurmikko TJ, et al. Control of chronic pain. BMJ November 21, 1998;317:1438-41.
Treatment of Lichen Planus: Evidence-Based Analysis
Main Published Trials of Lichen Planus Therapy
Number of studies by classification*
Treatment
A
B
C
Cutaneous lichen planus PUVA therapy Conventional 0 0 2 Bath 0 0 6 Etretinate (cutaneous) 0 0 3 Acitretin (cutaneous) 0 1 1 Oral tretinoin 0 0 4 Corticosteroids 0 0 1 Griseofulvin 0 0 4 Oral cyclosporine 0 0 4 Dapsone 0 0 1 Phenytoin 0 0 1 Mucosal lichen planus Topical corticosteroids 0 2 7 Systemic corticosteroids 0 0 2 Etretinate 0 0 6 Oral tretinoin 0 0 3 Topical tretinoin 0 0 5 Topical isotretinoin 0 0 2 Topical cyclosporine 0 0 18 Oral PUVA therapy 0 0 5 Extracorporeal photochemotherapy 0 0 1 Griseofulvin 0 0 3 Hydroxychloroquine 0 0 1
PUVA=psoralen plus ultraviolet A.
*--Classification according to Sackett's criteria. Anecdotal reports are not included.
Reprinted with permission from Cribier B, Frances C, Chosidow O. Treatment of lichen planus. An evidence-based medicine analysis of efficacy. Arch Dermatol 1998;134:1521-30.
Although lichen planus is a well-characterized dermatologic condition affecting skin, mucosa, hair and nails, treatment is often disappointing and even controversial. The varied manifestations of lichen planus result in markedly different clinical courses, which makes treatment planning extremely challenging. Spontaneous remissions occur more frequently with cutaneous lichen planus than with oral lichen planus. The mean duration of oral lichen planus is about five years, but the erosive form does not spontaneously resolve. Faced with the reported clinical variances of lichen planus, Cribier and associates performed a database review to evaluate treatment recommendations.
There are no large prospective trials with definitive results of the efficacy of various drug and treatment regimens. The largest controlled series included 65 patients. The authors selected 83 clinical trials and also analyzed case reports and review articles. The cutaneous and oral forms of lichen planus were examined separately. The criteria defined by Sackett were applied to establish the level of proof of effectiveness (levels A, B and C). Level A indicates large randomized controlled trials that allow definitive conclusions. Clinical trials with rigorous methods in which small numbers of patients were included are classified as level B. Controlled trials with less than 20 patients in each group were classified as level C, as well as trials without randomized controls.
The accompanying table summarizes the main published results. The authors' review demonstrated that there were no level A trials. The remainder of the controlled trials displayed disappointing results because of various methodological deficiencies, such as low numbers of subjects, faulty analysis of data or observational retrospective data. Many of the studies lacked precise clinical data and none of the studies used a quality-of-life scale; therefore, meta-analysis wasn't possible. The authors chose to consider only controlled studies in an attempt to define therapeutic indications using evidence-based analysis.
The first-line therapy in cutaneous lichen planus is acitretin. All other treatment methods or drugs are of uncertain efficacy. Based on clinical experience, systemic corticosteroids are recommended by many authors and could be classified as second-line treatment for cutaneous lichen planus. All other treatments, mainly psoralen followed by ultraviolet A (PUVA) light therapy and griseofulvin, need to be studied more rigorously and are not recommended at this time.
The first-line therapy in oral lichen planus (accepted in most reviews) is topical corticosteroids. No other therapy demonstrated convincing superiority over these agents. Second-line therapy in plaque-like lichen planus should be topical retinoids or etretinate; however, strong evidence in their support is lacking. All other therapies are unapproved or of uncertain or doubtful efficacy. In severe, multipledrug-resistant cases, topical cyclosporine could be recommended as a third-line treatment.
The authors conclude that their review demonstrated a lack of clear-cut results in the treatment of lichen planus, even for drugs that have been considered standard for some time. The promising newer therapies, such as topical cyclosporine, extracorporeal photochemotherapy or retinoids plus PUVA therapy, need to be tested in large controlled studies before widespread recommendations can be made.
BARBARA APGAR, M.D., M.S.
Cribier B, et al. Treatment of lichen planus. An evidence-based medicine analysis of efficacy. Arch Dermatol December 1998;134:1521-30.
EDITOR'S NOTE: This article was part of a series presented in the Archives of Dermatology addressing evidence-based analysis of common dermatologic therapies. This particular analysis of the literature points out how difficult it is to base clinical management on definitive data based on efficacy of therapy. Even though lichen planus is not the most common dermatologic disease, it may be associated with substantial morbidity and altered quality of life. After withdrawal of drug therapy, the recurrence rate is substantial.
B.A.
Once-Daily Amoxicillin for Streptococcal Pharyngitis
Four antibiotics--azithromycin, cefadroxil, cefixime and ceftibuten--are currently approved for once-daily administration in the treatment of group A beta-hemolytic streptococcal pharyngitis. These agents have a broad spectrum of activity and are expensive. The treatment of choice for streptococcal pharyngitis is penicillin, but it is ineffective in a once-daily regimen. Feder and associates conducted a randomized trial to determine if once-daily amoxicillin eradicates group A streptococcal pharyngitis. This regimen was compared to penicillin V given three times daily.
The 152 children in the study were between three and 18 years of age and presented with fever, tonsillar exudate, enlarged cervical lymph nodes and sore throat. Two swab specimens were obtained, one for a rapid antigen detection test and the other for a standard throat culture. Patients were randomly assigned to receive a 10-day course of either 250 mg of penicillin V three times daily or 750 mg of amoxicillin once daily. The first follow-up visit occurred within 24 hours of the initial visit.
Other follow-up visits occurred between the fourth and sixth days of therapy and between two weeks and three weeks after completing the assigned antibiotic therapy. Throat cultures were obtained at each follow-up visit. Patients receiving amoxicillin were switched to penicillin if any throat culture was positive for streptococcal organisms. Patients on penicillin who had positive throat cultures at any of the follow-up visits were given a second 10-day course of penicillin.
At the first follow-up visit, clinical response was comparable in the two treatment groups, with about 90 percent of each group clinically improved and afebrile. One patient in the penicillin group had a positive throat culture. No patient in the amoxicillin group had a positive culture.
At subsequent follow-up visits, 13 (16 percent) of the 79 patients in the amoxicillin group and 15 (21 percent) of the 73 patients in the penicillin group had positive throat cultures. In nine patients receiving amoxicillin and seven patients receiving penicillin, the organism was a a different strain of group A beta-hemolytic streptococci, and infection in these patients was considered a "new acquisition." The remaining four patients in the amoxicillin group and the remaining eight patients in the penicillin group were considered to have treatment failure because the organism was the same as that found initially.
The authors conclude that once-daily amoxicillin in a dosage of 750 mg is as effective in the treatment of streptococcal pharyngitis as penicillin V in a dosage of 250 mg three times daily. An advantage of amoxicillin is that its absorption is not affected by food intake (unlike penicillin) and its spectrum of activity is narrower than that of the antimicrobials currently approved for once-daily administration in the treatment of streptococcal pharyngitis. If other studies corroborate the findings of this study, amoxicillin may become another option for once-daily antibiotic therapy for streptococcal pharyngitis.
GRACE BROOKE HUFFMAN, M.D.
Feder HM, et al. Once-daily therapy for streptococcal pharyngitis with amoxicillin. Pediatrics January 1999; 103:47-51.
Methotrexate vs. Salpingostomy for Ectopic Pregnancy
In recent years, intramuscular methotrexate has gained attention as an alternative to salpingostomy for management of ectopic pregnancy. In addition, there is also interest in using serum human chorionic gonadotropin (hCG) or progesterone levels to monitor resolution of ectopic pregnancy after intervention. Saraj and associates compared methotrexate and salpingostomy in the treatment of ectopic pregnancy and evaluated the role of serum hCG and progesterone levels as markers for resolution.
The study included 75 patients with ectopic pregnancy; 37 underwent laparoscopic salpingostomy and 38 received an intramuscular injection of methotrexate (1 mg per kg). Patients were eligible for the study if the hCG level was greater than 2,000 mIU per mL, no gestational sac was visualized on ultrasound examination and the adnexal mass was less than 3.5 cm in size. Progesterone and hCG levels were determined at baseline, on days 4 and 7 after treatment, and weekly thereafter until the hCG level fell to less than 15 mIU per mL and the progesterone level declined to less than 1.5 ng per mL (4.5 nmol per L). A second methotrexate injection was administered if the hCG level on day 7 had not decreased by 15 percent from the level on day 4. Six (16 percent) of the 38 patients required a repeat injection.
Serum hCG levels in patients receiving methotrexate decreased to less than 15 mIU per mL in 27.2 ± 2.3 days, compared with 20.2 ± 2.7 days in patients undergoing salpingostomy. Initially, hCG levels increased from 3,162 ± 772 mIU per mL at baseline to 3,547 ± 808 mIU per mL four days after methotrexate injection. In contrast, hCG levels sharply decreased in the surgical group, from 3,356 ± 776 mIU per mL at baseline to 276 ± 90 mIU per mL four days after surgery.
In both treatment groups, serum progesterone levels decreased more rapidly than the hCG levels. They returned to less than 1.5 ng per mL in 17.6 ± 2.2 days in the methotrexate group and in 7.8 ± 1.7 days in the surgery group. Unlike serum hCG levels, serum progesterone levels did not show an increase on the fourth day after treatment.
Success rates were similar in the two groups: 95 percent (36 of 38) of the patients with methotrexate and 91 percent (33 of 36) of the patients with salpingostomy. Ipsilateral tubal patency rates three months after ectopic pregnancy were similar with either treatment, as were pregnancy rates among patients who were trying to conceive. Nine months after treatment, pregnancy had occurred in five (28 percent) of 18 patients treated with methotrexate and in four (29 percent) of 14 patients who had undergone salpingostomy.
The authors conclude that a single dose of intramuscular methotrexate is comparable to laparoscopic salpingostomy for the treatment of a small, unruptured ectopic pregnancy. That serum progesterone levels resolved faster than serum hCG levels suggests that serum progesterone may be a better marker for monitoring resolution of ectopic pregnancy.
BARBARA APGAR, M.D., M.S.
Saraj AJ, et al. Resolution of hormonal markers of ectopic gestation: a randomized trial comparing single-dose intramuscular methotrexate with salpingostomy. Obstet Gynecol December 1998:92:989-94.
Clinical Practices for Assessing Serum Digoxin Toxicity
The diagnosis of digoxin toxicity remains a clinical challenge, in part because therapeutic and toxic digoxin concentrations overlap from individual to individual. A key factor in determining if digoxin toxicity is present is whether or not blood is drawn at least six hours after the last digoxin dose, which assures that adequate distribution of the drug has been achieved before blood sampling. Williamson and associates evaluated the frequency of clinically defined digoxin toxicity in patients who were found to have elevated serum digoxin concentrations (i.e., greater than 2.0 ng per mL [2.6 nmol per L]).
The authors prospectively reviewed the medical records of adult patients in whom serum digoxin concentrations were determined at one of five hospitals during a three-month period. Data on the indications for digoxin therapy, evidence of toxicity and timing of the blood samples relative to administration of the last digoxin dose were collected for patients with serum digoxin concentrations greater than 2.0 ng per mL. A total of 3,434 serum digoxin determinations were obtained in 2,009 patients during the study period. Elevated levels occurred in 320 (9.3 percent) samples.
For 51 (15.9 percent) of these 320 elevated levels, blood was drawn six hours or less after administration of the last digoxin dose. For 70 assays, the sampling time could not be determined. Thus, the timing of blood sampling was appropriate in 199 (62.2 percent) of the 320 instances of elevated serum digoxin concentration.
The 199 appropriately timed blood samples were drawn from 138 patients, of whom 83 had at least one sign, symptom or electrocardiographic change suggestive of digoxin toxicity. The remaining 55 patients were asymptomatic and classified as having "no toxicity." Thus, the overall incidence of digoxin toxicity in this study population was 4.1 percent (83 of 2,009 patients); the authors classified 33 of the 83 patients as having "definite" digoxin toxicity and 50 as having "possible" digoxin toxicity. Cases identified as definite toxicity had a significantly higher mean serum digoxin concentration than those classified as possible toxicity or no toxicity.
Nausea was the most common symptom associated with digoxin toxicity. Patients identified as having definite toxicity had significantly more episodes of anorexia and diarrhea than did those considered to have possible toxicity. Heart block, bradycardia and junctional tachycardia were significantly more frequent in cases of definite toxicity compared with cases of possible toxicity.
A diagnosis of digoxin toxicity was documented in the medical records of 20 (61 percent) of the 33 patients with definite toxicity. Death occurred with equal frequency regardless of the presence or absence of digoxin toxicity.
The authors conclude that the findings of their study suggest a need for improvement in monitoring serum digoxin concentration. A sizable proportion of the assays in this study population were obtained without regard to digoxin pharmacokinetics. Such inappropriate timing can produce uninterpretable determinations, which can be costly for the hospital and the patient. With respect to determining serum digoxin concentrations, the authors suggest that the time at which blood is drawn and the time of the most recent digoxin dose be documented on all laboratory requisitions for serum digoxin levels. In addition, the authors recommend that laboratories not process blood samples drawn within six hours of the previous digoxin dose unless digoxin toxicity is strongly suggested on clinical grounds.
BARBARA APGAR, M.D., M.S.
Williamson KM, et al. Digoxin toxicity: an evaluation in current clinical practice. Arch Intern Med December 1998;158:2444-9.
Superior Vena Caval Filters for Upper Extremity DVT
Percutaneous placement of a filter in the inferior vena cava prevents pulmonary embolism from a deep venous thrombosis (DVT) in a lower extremity. Similarly, a filter in the superior vena cava may prevent pulmonary embolism from an upper extremity thrombus. However, placement of a filter into the superior vena cava is more challenging than insertion into the inferior vena cava because of the relatively small area for filter deployment. Theoretically, the technical demands of placing a filter in the superior vena cava could contribute to an increased complication rate. Spence and associates evaluated the safety and effectiveness of superior vena caval filters for the prevention of pulmonary embolism in 41 patients with upper extremity DVT.
Vena caval filters were required in these patients because of complications from or contraindications to anticoagulation. The patients were seen during a nine-year period. In no patient was the upper extremity DVT the reason for hospital admission. Rather, DVT developed in all of them following hospital admission for other conditions, including carcinoma (15 patients), congestive heart failure or recent myocardial infarction (11 patients), intracranial hemorrhage (seven patients), pancreatitis (two patients), elective surgery for arteriovenous malformation (two patients), gastrointestinal hemorrhage (one patient) and Henoch-Schöenlein purpura (one patient).
The extent of the upper extremity DVT was determined by venogram, duplex ultrasound examination, or both. The filters were placed, whenever possible, by way of the right common femoral vein to avoid inadvertent dislodgement of a central thrombus during internal jugular venous insertion.
In 22 of the patients, the thrombus was present at the site of a central venous catheter. In 14 of the patients, the thrombus was located where a central venous cathether had been located within the previous two weeks. Nine patients had hypercoagulable states. Sixteen of the patients had upper and lower extremity DVT simultaneously; 15 of these 16 patients underwent insertion of filters into both the superior and the inferior vena cavas.
No patient had immediate or delayed complications directly related to the procedure. At a median follow-up of 12 weeks (range: one day to 221 weeks), no evidence of migration, fracture or dislodgement of the filter was found in any patient, despite placement of central or Swan-Ganz catheters in 23 patients. No patient showed clinical evidence of superior vena cava occlusion, venous gangrene or exacerbation of upper extremity symptoms during a median follow-up of 15 weeks. One patient subsequently had a pulmonary embolism 44 months after placement of the superior vena caval filter. The clot was thought to have originated in the lower extremity. In 10 patients, anticoagulation was reinstituted while the filter was in place, and the filter likely provided additional protection against pulmonary embolism.
In the authors' experience, mechanical interruption of the superior vena cava was 100 percent effective in preventing symptomatic pulmonary embolism related to upper extremity DVT. Although the survival rate at six months was low (48 percent) in this study group, deaths were largely from cancer and cardiac disease.
BARBARA APGAR, M.D., M.S.
Spence LD, et al. Acute upper extremity deep venous thrombosis: safety and effectiveness of superior vena caval filters. Radiology January 1999;210:53-8.
Diagnosing PID: Comparing Ultrasound, MRI, Laparoscopy
Laparoscopy is considered the standard of care in the diagnosis of pelvic inflammatory disease (PID), but it has the disadvantage of requiring general anesthesia. Transvaginal ultrasonography, which is superior to abdominal ultrasonography in demonstrating evidence of PID, requires the presence of thickened, fluid-filled fallopian tubes to make the diagnosis of PID. Although computed tomography (CT) can identify complex tubo-ovarian abscesses and has proved useful in the diagnosis of PID, it has the disadvantage of exposing reproductive organs to ionizing radiation. Magnetic resonance imaging (MRI) may provide better images of the soft tissue than CT. Tukeva and associates evaluated the accuracy of MRI in the diagnosis of PID and compared its accuracy with that of transvaginal ultrasonography and laparoscopy.
The study included 30 consecutive patients admitted to the hospital because of symptoms of PID. Each patient was evaluated by means of transvaginal ultrasonography, followed by MRI 24 hours later. Laparoscopy was performed immediately after MRI. If an adnexal mass was present, surgical samples were obtained for histology. Laparoscopy confirmed PID in 21 (70 percent) of the 30 patients. Three of the nine patients without PID had tubal torsion. In the other patients without PID, laparoscopy revealed a simple cyst, a dermoid cyst, an endometrioma, a ruptured cyst and free pelvic fluid. One patient had no laparoscopic evidence of a gynecologic disorder.
MRI findings were consistent with the diagnosis of PID in 20 (95 percent) of the 21 patients with laparoscopically proven PID and in one of the nine patients without PID. MRI demonstrated an abscess in 11 of the 21 patients with PID. False-negative MRI findings occurred in only one patient. In this patient, MRI was thought to show an endometrioma and a hemorrhagic cyst; laparoscopy revealed an endometrioma and salpingitis. Conversely, false-positive MRI findings occurred in one patient; this patient was thought to have tubal torsion and pyosalpinx on MRI, but laparoscopy revealed tubal torsion and hydrosalpinx.
Transvaginal ultrasonographic findings were consistent with the diagnosis of PID in 17 (81 percent) of the 21 patients with laparoscopically proven PID and in two of the nine patients without PID. Of the latter two patients, one had an endometrioma and the other had tubal torsion. Three cases of PID were missed by transvaginal ultrasonography. Two of these patients were thought to have an ovarian tumor on ultrasonographic examination but at laparoscopy were found to have an abscess. In one patient, endometrioma was diagnosed by ultrasonography, but this patient was found at laparoscopy to have pyosalpinx in addition to endometrioma.
Overall, MRI had a sensitivity of 95 percent, a specificity of 89 percent and an accuracy of 93 percent in the diagnosis of PID. For transvaginal ultrasonography, the sensitivity, specificity and accuracy rates were 81 percent, 78 percent and 80 percent, respectively.
The authors conclude that MRI was useful not only for establishing the diagnosis of PID but also for detecting other processes responsible for the symptoms. Although it could be argued that transvaginal ultrasonography is far more cost-effective than MRI, the authors state that if proper therapy is initiated early in the disease course and if exploratory laparoscopy can be avoided on the basis of the MRI findings, then MRI could prove to be a cost-effective modality.
BARBARA APGAR, M.D., M.S.
Tukeva TA, et al. MR imaging in pelvic inflammatory disease: comparison with laparoscopy and US. Radiology January 1999;210:209-16.
EDITOR'S NOTE: The difficulty in accurately diagnosing PID has prompted the Centers for Disease Control and Prevention (CDC) to recommend that antibiotic therapy be initiated in any patient presenting with adnexal, uterine and cervical tenderness. According to the CDC, these signs are sufficient evidence of PID to justify antibiotic therapy. If no response occurs in 72 hours, the patient should be reevaluated, and laparoscopy should be considered. Whether use of MRI can help decrease the need for laparoscopy in patients suspected of having PID requires further study. Transvaginal ultrasonography has not been shown to be accurate as the sole imaging method for the diagnosis of PID, although it is widely used in the evaluation of patients presenting with pelvic pain suggestive of PID.
B.A.
Benefit of Beta Blockers in Patients with Heart Failure
Although beta-blocking medications have been demonstrated to improve left ventricular function and morbidity in patients with heart failure, concerns about safety and the overall effect on mortality have limited the routine use of these agents. Meta-analysis of several clinical trials has shown an estimated reduction in mortality of about 30 percent. A large European multicenter trial reports on the effect of beta blockade on rates of mortality in patients with heart failure.
The Cardiac Insufficiency Bisoprolol Study group included over 100 investigators in 20 countries, and 2,647 symptomatic patients with heart failure were enrolled. Patients were 18 to 80 years of age with symptoms consistent with New York Heart Association (NYHA) class III or IV heart failure and left ventricular ejection fractions of 35 percent or less. All patients had been diagnosed at least three months previously. Treatment was required to include diuretics and angiotensin converting enzyme (ACE) inhibitors, although use of digoxin and other medications was optional. Patients were randomly assigned to receive 1.25 mg of bisoprolol or an identical placebo daily. The dosage was gradually increased to 10 mg daily as tolerated. Patients were monitored every three months by study personnel.
The study was stopped early because the mortality rate in the bisoprolol patients was significantly less than that in patients receiving placebo. The mean follow-up time was 1.3 years. During this period, 156 treated patients (11.8 percent) died, compared with 228 patients in the placebo group (17.3 percent). Patients receiving bisoprolol were significantly less likely to die from or be admitted to the hospital for a cardiovascular event. These differences were maintained when data were analyzed by cause and severity of heart failure. The incidence of sudden death was reduced by 42 percent in patients receiving bisoprolol (48 sudden deaths compared with 83 in the placebo group). Hospital admission for deteriorating heart failure was reduced by 32 percent (159 patients in the treatment group compared with 232 in the placebo group). Admissions were also significantly reduced for ventricular tachycardia, ventricular fibrillation and hypotension but were increased for stroke, although the increase was not significant.
The authors conclude that beta blockade has a significant benefit in patients with NYHA class III and IV heart failure. These findings confirm the benefit projected from meta-analyses and other studies. They recommend that beta blockers be added to standard therapeutic regimens of diuretics and ACE inhibitors in patients with stable heart failure. Since much remains to be elucidated about the optimal use of these combinations of treatments, they recommend starting with a low dosage of a beta blocker and progressively increasing to the maximum tolerated dosage. In this study, 10 mg daily was the most common maintenance dosage.
ANNE D. WALLING, M.D.
CIBIS-II Investigators and Committees. The Cardiac Insufficiency Bisoprolol Study II (CIBIS-II): a randomised trial. Lancet January 2, 1999;353:9-13.
Reduction in Mortality from CHD with Use of Pravastatin
The epidemiologic relationship between serum cholesterol levels and the risk of coronary heart disease (CHD) is well documented. Treatments designed to lower serum cholesterol levels in patients with moderately high cholesterol have had limited effect. Patients receiving cholesterol-lowering therapy experience fewer coronary events, but overall coronary mortality is reduced by only about 10 percent. Because of the slight increase in deaths from noncoronary causes, the effects of cholesterol-lowering therapy on overall mortality in patients with high cholesterol levels, and the risk of coronary events in patients with lower cholesterol levels remain uncertain. The Long-term Intervention with Pravastatin in Ischaemic Disease (LIPID) study group conducted a double-blind, randomized trial to study the effects of lowering cholesterol levels with the HMG-CoA reductase inhibitor pravastatin in patients with a history of myocardial infarction or unstable angina.
The LIPID group recruited patients aged 31 to 75 years from 87 sites in Australia and New Zealand. Patients were eligible for the study if they had a hospital discharge diagnosis of myocardial infarction or unstable angina in the previous three to 36 months. Plasma cholesterol levels had to be from 155 to 271 mg per dL (4.0 to 7.0 mmol per L). Fasting triglyceride levels had to be less than 445 mg per dL (5.0 mmol per L).
During an initial eight-week placebo period, patients received dietary instruction aimed at reducing fat intake to less than 30 percent of the daily caloric intake. Patients were then given either 40 mg of pravastatin or placebo daily, along with continued dietary advice. Cholesterol levels were obtained at randomization, after six months and then annually until the end of the six-year study period. High-density lipoprotein (HDL), low-density lipoprotein (LDL) and triglyceride levels were also obtained, but at slightly different intervals.
The primary outcome of the study was death from CHD. Secondary outcomes included death from any cause, stroke, need for heart surgery, number of days in the hospital and reduction in serum lipid levels.
Between 1990 and 1992, a total of 9,014 patients were randomly assigned to receive either 40 mg of pravastatin or placebo daily. Baseline characteristics were similar between the groups. Patients were predominantly male and had a median age of 62 years. The median cholesterol level in both groups was 218 mg per dL (5.6 mmol per L). Reductions in total and LDL cholesterol in the group receiving pravastatin were 18 and 25 percent greater than in the placebo group. By the end of the study, 19 percent of the patients in the treatment group had stopped taking pravastatin, and 24 percent of patients in the placebo group had crossed over to therapy with a cholesterol-lowering drug.
The effect of treatment with pravastatin was encouraging. The incidence of death from CHD was 6.4 percent in patients receiving pravastatin, compared with 8.3 percent in patients receiving placebo, for a relative risk reduction of 24 percent. Overall mortality decreased by 22 percent in the pravastatin group, and mortality from cardiovascular causes decreased by 25 percent. In addition, the incidence of myocardial infarction and stroke, the rate of bypass surgery and angioplasty, and the rate of hospitalization and number of hospitalized days were also lower for patients in the treatment group.
The authors conclude that in patients with CHD, using pravastatin to lower cholesterol levels reduces mortality rates from cardiovascular disease and all causes combined. The risk of myocardial infarction and stroke is also reduced. Pravastatin appears to be safe and well tolerated, as the authors found no significant increases in adverse events associated with taking the medication. At present, approximately 30 percent of patients are given lipid-lowering drugs following acute myocardial infarction. Given these results, the authors recommend that treatment with lipid-lowering drugs be considered in all patients with CHD.
JEFFREY T. KIRCHNER, D.O.
The Long-term Intervention with Pravastatin in Ischaemic Disease (LIPID) Study Group. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. N Engl J Med November 5, 1998;339:1347-57.
Should Stress Test Be Routine in Patients with Chest Pain?
Of the more than 5 million persons who present to emergency departments each year with chest discomfort suggesting myocardial ischemia, 75 percent or more have no objective evidence of an unstable coronary syndrome. Evaluation of these "low-risk" patients with hospitalization is very expensive, necessitating the development of rapid "rule-out" protocols in short-stay observation areas to reduce costs. Lindsay and associates point to the need for an effective and accurate protocol in these short-stay situations and examine some of the data that support a protocol calling for routine stress testing.
Virtually all triage protocols include a period of observation for recurrent ischemic pain and electrocardiographic (ECG) changes, as well as testing for serum markers of myocardial necrosis. Some protocols include routine stress testing if there is no evidence confirming active ischemia during the observation period.
The stress test most often used, exercise ECG, has important disadvantages as a screening test in acute ischemia triage situations. First, many patients cannot exercise adequately for reasons other than heart disease. Second, exercise ECG results can be misleading when the resting ECG is abnormal. Patients with left ventricular hypertrophy, left bundle branch block or resting ST-T wave segment abnormalities cannot be accurately evaluated. The false-positive rate in a low-risk population such as those with negative findings after initial screening for unstable coronary disease is unacceptably high.
Echocardiography, stress echocardiography and myocardial scintigraphy have all been advocated as ways to recognize patients with unstable coronary artery disease. These studies provide higher sensitivity and specificity than exercise ECG and can be used in more diverse patients. Although these procedures are more accurate and improve the triage procedure, they increase the cost of chest screening tremendously.
The authors concur with the view of the National Heart Attack Alert Coordinating Committee's Working Group on Evaluation of Technologies for Identifying Acute Cardiac Ischemia in the Emergency Department, which found that data supporting the routine use of stress testing to be limited and inconclusive. Newer biochemical assays including troponin levels combined with clinical markers such as age, history of prior coronary artery disease and the presence of cardiac risk factors such as diabetes may be the most effective way to triage patients with chest pain.
RICHARD SADOVSKY, M.D.
Lindsay J, et al. Routine stress testing for triage of patients with chest pain: is it worth the candle? Ann Emerg Med November 1998;32:600-3.
Hyperlipidemia in Patients with Type 2 Diabetes
In patients with diabetes, coronary artery disease is the most common cause of death. Lipid abnormalities are commonly associated with diabetes, particularly in those with type 2 diabetes (formerly known as noninsulin-dependent diabetes). The most common lipid abnormalities in these patients include hypertriglyceridemia and reduced high-density lipoprotein (HDL) cholesterol levels. While lipid abnormalities typically improve with better glycemic control, normalization does not usually occur. Because there is a strong relationship between all forms of vascular disease in patients with type 2 diabetes and hyperlipidemia, it is important to screen for and treat these lipid abnormalities. O'Brien and associates review recent studies of the evaluation and management of this problem.
Annual screening for lipid abnormalities in adults with diabetes is recommended. Such screening should include measurements of total cholesterol, HDL, low-density lipoprotein (LDL) and triglyceride levels. An acceptable LDL level is less than 130 mg per dL (3.35 mmol per L); triglycerides should be less than 200 mg per dL (2.25 mmol per L). In patients with clinically evident vascular disease, LDL levels should be less than 100 mg per dL (2.60 mmol per L), and triglycerides should be less than 150 mg per dL (1.70 mmol per L). Whether these lower values should be the target for all patients with diabetes, regardless of whether they manifest vascular disease, has been debated. An HDL level of greater than 45 mg per dL is recommended (1.15 mmol per L).
Management of hyperlipidemia should begin with improving glycemic control and losing weight. Exercise should be incorporated into a weight-loss program, as it has been shown to enhance weight loss and facilitate weight maintenance. Weight loss will result in a decrease in triglyceride levels and an increase in HDL levels. Before an exercise program can be recommended, concomitant medical conditions that would increase the risks of exercise should be taken into consideration, including the presence of proliferative retinopathy, neuropathy and foot problems. It is prudent to recommend an exercise tolerance test to rule out silent myocardial ischemia, particularly in patients older than 35 years.
If the goals for lipid levels have not been reached after three to six months of diet, exercise and improved glycemic control, drug therapy should be initiated. However, drug therapy should be used at the outset in patients with severe hypertriglyceridemia (triglyceride level greater than 1,000 mg per dL [11.30 mmol per L]). The type of drug chosen should be based on the lipid abnormality that is present. In patients with hypercholesterolemia without hypertriglyceridemia, an HMG-CoA reductase inhibitor should be used; in patients with hypercholesterolemia with hypertriglyceridemia, an HMG-CoA reductase inhibitor or gemfibrizol can be used; in patients with hypertriglyceridemia, gemfibrizol can be used. A patient with decreased HDL levels may benefit from taking an HMG-CoA reductase inhibitor or niacin; however, niacin should be used with caution because of a possible adverse effect on glycemic control. Omega-3 fatty acids (fish oils) have been shown to reduce lipid levels in healthy patients. However, when fish oils have been used in patients with type 2 diabetes, some adverse effects have been reported, including elevation of fasting and postprandial glucose levels. Combination drug therapy can be used if hyperlipidemia is unresponsive to monotherapy. An extremely useful example is the combination of low-dose bile acid sequestrants and HMG-CoA reductase inhibitors. The use of an HMG-CoA reductase inhibitor with fibrates or niacin is associated with an increased risk of myopathy. While not contraindicated, this combination should be used with caution.
The authors conclude that hyperlipidemia is partly responsible for the increased vascular disease that occurs in patients with diabetes. Hypertriglyceridemia and reduced LDL levels should be aggressively managed in these patients. Effective treatment includes a combination of pharmacologic and nonpharmacologic therapy. All adults with diabetes should receive an annual fasting lipoprotein profile.
JIM NUOVO, M.D.
O'Brien T, et al. Hyperlipidemia and diabetes mellitus. Mayo Clin Proc October 1998;73:969-76.
Can Lifestyle Changes Be an Option to Revascularization?
Experience has shown that progression of coronary artery disease can be arrested and reversed by changes in diet and lifestyle. Because of the high cost of treatment of coronary artery disease in the United States (an estimated $56.3 billion in 1994), there is considerable incentive to finding alternative interventions that are less expensive, yet effective. Ornish describes the Multicenter Lifestyle Demonstration Project, a one-year program that was designed to determine if comprehensive lifestyle changes could be a cost-effective alternative to revascularization in select patients with severe but stable coronary artery disease.
The program proposed lifestyle changes that included a very low-fat, low-cholesterol diet (whole-foods vegetarian diet, approximately 10 percent fat, less than 10 mg per day dietary cholesterol, high in complex carbohydrates and low in simple sugars), stress management techniques, moderate exercise and psychosocial support.
This program was offered as a less-costly alternative treatment to revascularization in selected patients eligible for bypass graft or angioplasty. A bypass graft is effective in decreasing angina and improving cardiac function. When compared with medical therapy, however, survival is improved in only a specific group of patients: those with decreased left ventricular function and stenotic lesions of the left main coronary artery of greater than 59 percent. The mortality and morbidity benefits of angioplasty are unknown because it has never been compared with medical therapy in stable patients with coronary artery disease.
Patients eligible for the program had angiographically documented coronary artery disease severe enough to warrant revascularization. Exclusion criteria included (1) greater than 50 percent stenosis in the left main coronary artery, (2) bypass graft within six weeks or angioplasty within six months, (3) chronic unresponsive congestive heart failure, (4) malignant uncontrolled arrhythmias, (5) myocardial infarction within one month, (6) homozygous hypercholesterolemia, (7) psychosis, (8) hypotensive response to exercise, (9) alcohol or drug abuse and (10) life-threatening co-morbidity.
The author notes that the lifestyle-change program required patients to have commitment, discipline and willingness to accept personal responsibility for their health. Those who decided to undergo revascularization more often wanted a "quick fix," making randomization for a controlled trial impossible. Of the 194 patients in the experimental group, 150 were able to avoid revascularization for at least three years, and the frequency of adverse cardiac events was not increased. Reductions of angina were also achieved in the experimental group that were comparable with those achieved with revascularization. The average cost saving per patient was almost $30,000.
The author concludes that patients with coronary artery disease should be offered several therapeutic options, including comprehensive lifestyle changes, medications (including lipid-lowering drugs), angioplasty and bypass surgery. Unfortunately, most third-party payers will cover the drug therapy and revascularization but not the cost of instructing patients in a lifestyle-change program. This lack of widespread insurance coverage is more of a limiting factor for lifestyle changes than is a shortage of motivated patients.
RICHARD SADOVSKY, M.D.
Ornish D. Avoiding revascularization with lifestyle changes: the multicenter lifestyle demonstration project. Am J Cardiol November 26, 1998;82:72T-6T.
EDITOR'S NOTE: Lifestyle modification requires a highly motivated patient and a clear treatment plan. Using techniques similar to those used in all treatments, Winslow and associates outlined three steps to accomplishing lifestyle changes in patients: (1) help the patient understand the value of the treatment, (2) discuss ways in which the treatment will evolve and set appropriate goals, and (3) monitor and encourage progress while identifying barriers or adverse effects (Am J Med 121:4A25S-31S, October 8, 1996). Programs need to be individualized, with the patient receiving frequent feedback and encouragement. Written material is a helpful adjunct to patient education and enthusiasm. Lifestyle modification may require significant investments in time but may help a patient avoid a treatment or condition with higher morbidity.
R.S.
Is Oral Treatment of Toenail Onychomycosis Effective?
It is well known by clinicians and patients that topical treatment of toenail onychomycosis, specifically with griseofulvin, is generally ineffective. When antifungal agents such as fluconazole, itraconazole and terbinafine were first introduced, reports of cure rates in the range of 35 to 80 percent began to appear. Epstein analyzed studies through Medline that purported to define the efficacy of oral treatment of toenail onychomycosis and also sought to determine how long the nail actually remained disease free.
Twenty-six articles that described clinical results, used both culture and microscopy for diagnosis and included clinical evaluation data were included in the study. However, only seven studies provided results in terms of a "disease-free nail." These studies covered 11 trials, three with itraconazole and eight with terbinafine.There has been concern as to how often and when a successful dermatologic result turns into a failed treatment. Depending on the study, such failures have been labeled "relapse," "recurrence" or "reinfection." The important aspect is the reappearance of nail dystrophy, irrespective of its mechanism. "Disease reappearance" describes a disease-free nail without implying the mechanism responsible.
Three months of treatment with itraconazole produced a disease-free nail in about 35 percent of patients, whether or not the treatment was given continuously at 200 mg per day or as a one-week pulse each month of 400 mg per day. Mycologically negative patients whose nails had cleared or had markedly improved were referred to as an "overall success" in one study. In that study, fewer than 35 percent of the patients would have achieved a disease-free nail according to the definition.
Treatment with terbinafine achieved a disease-free nail in about 40 to 50 percent of patients, except in one study (of 17 subjects), which reported a 76 percent cure rate. Although the regimen that is usually recommended for terbinafine treatment of toenails is 250 mg daily for 12 weeks, in five of the eight trials it was prescribed for 16 weeks or longer. Only one study adequately assessed the frequency of disease reappearance. Of 31 patients who achieved a disease-free nail at one year of follow-up, five patients showed recurrent nail dystrophy at the two-year evaluation, resulting in a reappearance rate of 17 percent. However, this result applied only to the appearance of the nail. The mycologic failure rate was much higher. At the two-year evaluation, only 55 percent of the fungal cultures were negative. Forty-two percent of patients demonstrated mycologic failure during a one-year period. Although itraconazole and terbinafine have been in clinical trials for more than eight years, only one study provided data at both one and two years after treatment.
Results of this analysis demonstrate the importance of determining the negativity of results of microscopy and fungal culture. The assumption behind this goal is that failure to clear the fungal infection will sooner or later result in a dystrophic nail. Many of the studies used the term "clinical remission" as their end points but failed to further define the term. Of the 26 studies reviewed, only 14 used a normal-appearing nail as the end point.
The author concludes that in order to successfully treat patients with toenail onychomycosis, physicians must have the data to know not only how often the infection can be cleared but how long the nails will remain free of disease. Disease reappearance is an important issue faced by clinicians attempting to treat toenail onychomycosis.
BARBARA APGAR, M.D., M.S.
Epstein E. How often does oral treatment of toenail onychomycosis produce a disease-free nail? An analysis of published data. Arch Dermatol December 1998;134: 1551-4 and Bigby M. Snake oil for the 21st century. Arch Dermatol December 1998;134:1512-4.
EDITOR'S NOTE: In an accompanying essay, Bigby stresses the importance of basing dermatologic therapy on the best evidence available, rather than on anecedotal or expert opinion. With the advent of free-standing clinics offering such destructive modalities as chemical peels, dermabrasion and cryotherapy, it is up to the referring physician to discuss new therapies honestly with the patient--are new therapies more effective than the existing ones that may not be as heavily or as shrewdly marketed? The editors of this dermatology journal should be commended for addressing the importance of an evidence-based approach to therapy.
B.A.
"Tips from Other Journals" are written by the medical editors of American Family Physician.
Copyright © 1999 by the American Academy of Family Physicians.
This content is owned by the AAFP. A person viewing it online may make one printout of the material and may use that printout only for his or her personal, non-commercial reference. This material may not otherwise be downloaded, copied, printed, stored, transmitted or reproduced in any medium, whether now known or later invented, except as authorized in writing by the AAFP.
