Human immunodeficiency virus disease continues to evolve very rapidly, with guidelines that require updating several times a year. New infections continue to emerge both in the United States and around the world. Biological weapons must now be considered in discussions of emerging infections. And finally, resistance to antibiotics is spreading rapidly, and almost no community remains unaffected. As a consequence, physicians treating infectious diseases at the end of the 20th century may face clinical travails similar to those of physicians who treated infections at the century's beginning.

General Infectious Diseases

Numerous current issues in general infectious diseases are intimately related. For example, the increase in resistance to antibiotics is largely due to antibiotic overuse. Infections involving antibiotic-resistant pathogens are emerging, and international travel has made this a global problem.

Antibiotics Were Still Commonly Used for Self-Limited Infections

Gonzales R, Steiner JF, Sande MA. Antibiotic prescribing for adults with colds, upper respiratory tract infections, and bronchitis by ambulatory care physicians. JAMA. 1997; 278:901-4.

Previous studies showed that upper respiratory tract infections and acute bronchitis are the reason for as many as 30% of all antibiotic prescriptions. Such unnecessary use of antibiotics is believed to contribute heavily to the growing resistance of Streptococcus pneumoniae to penicillin. For example, the prevalence of carriage of resistant S. pneumoniae in day care centers has increased to 29% [1].

Gonzales and colleagues sought to measure prescribing rates and identify predictors of antibiotic use for adults who received diagnoses of colds, upper respiratory tract infections, and acute bronchitis. They identified 1529 practicing physicians who participated in the National Ambulatory Medical Care Survey in 1992, and they obtained information on diagnoses and prescriptions for 28 787 office visits. Patients with underlying lung diseases were excluded from the analyses. From these samples, the investigators estimated practice patterns throughout the United States.

Collectively, diagnoses of bronchitis, upper respiratory tract infection, and pharyngitis were the reason for 30% of all antibiotic prescriptions. Furthermore, more than half of visits for these diagnoses resulted in antibiotic prescriptions (Figure 1).

View larger version (18K): [in this window] [in a new window]   Figure 1. Number of office visits (striped bars) and antibiotic prescriptions (white bars) for common infections in adults: result of the National Ambulatory Care Survey, 1992. Data from Gonzales R, Steiner JF, Sande MA. Antibiotic prescribing for adults with colds, upper respiratory tract infections, and bronchitis by ambulatory care physicians. JAMA. 1997; 278:901-4. URI = upper respiratory tract infection; UTI = urinary tract infection.

These data are not surprising, but they are very worrisome, and the Centers for Disease Control and Prevention (CDC) has been concerned enough to establish guidelines for the use of antibiotics for common infections. These guidelines have been published for pediatric practice and will soon be published for adults [2-5]. The adult guidelines are expected to reflect those for pediatric practice, so a brief review of the latter should help guide adult practice at this time.

Because group A Streptococcus infections account for only 15% of cases of pharyngitis, the CDC recommends that patients be treated only for laboratory-confirmed group A streptococcal infections and that penicillin be the drug of choice. Clinical criteria-fever, exudate, and lymphadenopathy-are insufficient justification for empirical antibiotic use in this era of growing antibiotic resistance. The CDC guideline indicates that positive results on either culture or the rapid antigen test comprise acceptable justification. But if the rapid antigen test result is negative, the decision to treat should be based on culture because the rapid test lacks sensitivity.

The common cold is one of the main triggers of antibiotic abuse. Acute bronchitis also results in numerous antibiotic prescriptions, even though acute bronchitis is almost always due to viral infection and six clinical trials in adults have shown that antibiotics have not demonstrated benefit in these cases [6]. One possible exception is bronchitis resulting from pertussis infection, which is a growing problem in adults in the United States (about 5000 cases were reported in 1997). However, if pertussis has been present for more than a week, antibiotics do not help eradicate symptoms (although they do reduce transmission). Other exceptions include mycoplasma infection and bronchitis in patients who are seriously immunocompromised or have severe lung disease.

Mucopurulent rhinitis as part of a head cold is usually associated with sinusitis and generally should not be treated with antibiotics. Antibiotics may be indicated in patients who have had symptoms for at least 10 days or in patients with facial swelling, pain, or body temperature greater than 39 °C. Gonzales and colleagues do not advocate radiologic studies because most patients with simple upper respiratory tract infections have evidence of sinusitis on computed tomography. Such studies might be helpful, however, if complications occur, if the diagnosis is in doubt, or if patients do not respond to antibiotics or other therapy. Amoxicillin remains the preferred initial antibiotic, and treatment should be given for 10 to 14 days. For patients who do not respond in 2 to 3 days, it might be reasonable to switch to alternative drugs, such as amoxicillin-clavulanate or a cephalosporin.

Of course, a major predictor of antibiotic use is patient preference. To address this issue, the CDC and the American Academy of Pediatrics are launching a campaign to inform the public about the dangers of the inappropriate use of antibiotics. This campaign may make it easier for clinicians to say "no" when antibiotics are not needed.

Avian Influenza Has Potential for Catastrophe

Yuen KY, Chan PK, Peiris M, et al. Clinical features and rapid viral diagnosis of human disease associated with avian influenza A H5N1 virus. Lancet. 1998; 351:467-71.

Longevity in the United States has increased steadily over the past century, except for the 11-year decrease during the 1918-1919 winter outbreak of Spanish influenza, which killed about 20 million persons. Since that time, two more major influenza epidemics have occurred, the 1958-1959 Asian influenza epidemic and the 1967-1968 Hong Kong influenza epidemic.

Public health experts fear another epidemic similar to the Spanish influenza epidemic, the virulence of which has never been explained. That fear increased in May 1997 with the report of the death, from influenza, of a 3-year-old boy in Hong Kong [7]. The implicated influenza strain turned out to be the first isolate of H5N1 in a human. This strain was identical to an avian strain of influenza that was simultaneously epidemic among chickens in Hong Kong. Eighteen persons eventually fell ill with this particular strain: Eight required mechanical ventilation, and 6 died. No human-to-human transmission could be shown by serologic testing of case contacts, poultry workers, and controls. The decision to slaughter all chickens in Hong Kong was made on 28 December 1997, and no further cases have been reported.

The report by Yuen and colleagues, from the Hong Kong Health Department, identified the clinical features of the first 12 cases and proposed a method of rapid diagnosis. Patients, who ranged in age from 1 to 60 years, had gastrointestinal symptoms, hepatitis, renal failure, and pancytopenia-evidence of a much more systemic illness than is suggested by the dominance of respiratory symptoms usually associated with influenza. Also in contrast to typical influenza, death occurred as a result of the primary viral infection rather than secondary bacterial infections.

To cause a global pandemic, an outbreak of influenza must meet three conditions. First, specific antibody must be absent in the general population. Second, the viral strain must be able to replicate in humans. And third, the viral strain must be transmissible from person to person. The strain seen in Hong Kong possessed the first two criteria but apparently not the third. It was assumed, however, that if the virus had remained active, it would have mutated and acquired transmissibility. This was the rationale for the decision to slaughter all chickens.

Ribavirin Was Added to Hepatitis C Therapy

Reichard O, Norkrans G, Fryden A, et al. Randomised, double-blind, placebo-controlled trial of interferon -2b with and without ribavirin for chronic hepatitis C. The Swedish Study Group. Lancet. 1998; 351:83-7.

Hepatitis C virus (HCV) is possibly responsible for more cases of cirrhosis than alcohol. It is thought to infect more than 4 million Americans, although fewer than 5% of them know that they are infected. Chronic infection develops in about 85% of infected persons, and cirrhosis occurs in roughly 20% of those with chronic infection after 20 years. Blood transfusions done before routine testing began account for about 5% of infections. Perhaps surprisingly, sexual and perinatal transmission of HCV seem to be relatively uncommon.

The only treatment available for hepatitis C has been interferon- 2b, which produces persistent benefit in only 10% to 20% of patients and does so at a cost of often severe adverse effects and more than $5000 for a course of treatment. Recently, several small studies have shown that ribavirin, a nucleoside reverse transcriptase inhibitor, plus interferon- 2b produced more sustained benefit in patients with chronic hepatitis C than did interferon alone.

In this clinical trial, 100 patients with chronic hepatitis C were randomly assigned to receive interferon- 2b (3 million U three times weekly) with either placebo or ribavirin (1000 or 1200 mg/d) for 24 weeks. All patients were followed for 24 additional weeks after therapy ended. Criteria for inclusion were persistent elevation of aminotransferase levels for at least 6 months, measurable levels of HCV RNA, and findings of chronic hepatitis on liver biopsy. Analyses were done on an intention-to-treat basis.

At 48 weeks, HCV RNA levels were undetectable in 46% of those receiving combined therapy and 21% of those receiving interferon alone. Overall, mean HCV RNA levels after treatment were 3.2 million copies/mL and 7.4 million copies/mL, respectively. Therapy was completed by 94% of those receiving interferon alone and 86% of those receiving combination therapy. As expected, anemia and nausea were common adverse effects of ribavirin.

Ribavirin may now become part of standard treatment for chronic hepatitis due to HCV. Whether the improved response rate will lead to improved overall long-term prognosis is unknown.

Cyclospora Species Brought Disease Internationally

Herwaldt BL, Ackers ML. An outbreak in 1996 of cyclosporiasis associated with imported raspberries. The Cyclospora Working Group. N Engl J Med. 1997; 336:1548-56.

In recent years, four epidemics of food-borne disease have received substantial attention. Salmonella organisms from an ice cream plant produced symptomatic salmonellosis in at least 220 000 persons in 1994 [8]. Strawberries contaminated with hepatitis A virus produced illness in 153 persons [9]. Escherichia coli O157 from beef affected only 15 persons, but the outbreak forced a food producer out of business [10].

The fourth recent outbreak was of infection with Cyclospora cayetanensis, a newly recognized spore-forming protozoan that causes gastroenteritis. In many ways, this organism resembles Cryptosporidia, Microsporidia, and Isospora species. This Cyclospora species causes a protracted or relapsing form of diarrhea that lasts for 2 to 3 weeks. Like Cryptosporidia species, it can be identified by acid-fast stain of the stool. Unlike Cryptosporidia species, it can be effectively treated with trimethoprim-sulfamethoxazole.

In a retrospective cohort study, Herwaldt and Ackers identified 1465 cases of cyclosporiasis that occurred in 20 states in 1996. About half of the cases (n = 725) were associated with 55 events, such as picnics, held between 3 May and 14 June 1996. Using epidemiologic surveys, the investigators found that raspberries had been served at 50 of the 55 events. Persons from a sample of the events were interviewed, and the association between consumption of raspberries and symptoms of Cyclospora infection was high (P < 0.05).

The raspberries came from five farms in Guatemala. When these farms were investigated, the Cyclospora species was not found but evidence of fecal contamination was present.

Unfortunately, before the study was completed, strawberries were implicated as the source of infection. This observation rapidly appeared in the media, and the effect on the California strawberry industry was devastating. It was later found that strawberries were often served with raspberries; this explained the erroneous implication of strawberries.

The Cyclospora outbreak highlights several difficult issues. The first concerns the identification and reporting of disease. We know that most enteric infections are never reported. For salmonellosis, which is easy to diagnose, we know that less than 5% of cases are reported. For pathogens such as Cyclospora species, most laboratories do not offer tests. The diagnosis may therefore remain unknown.

Second, the source of the oocytes is unclear. One particular problem is that raspberries, unlike fruits such as apples, cannot be washed effectively. Understanding how the oocytes are spread will be necessary to prevent the fruit from becoming contaminated in the first place.

Finally, and perhaps most important, contamination of fresh fruit is increasingly an international issue. Until recently, fresh food in the United States and elsewhere came from regional supplies. But fresh food distribution has become so efficient that pathogens can spread rapidly over wide areas of the world. This mass distribution is responsible for the magnitude of many recent outbreaks. Several methods for controlling the outbreaks have been proposed; among them, food irradiation is clearly gaining appeal.

Antibiotics Resulted in Fewer Cardiac Events

Gurfinkel E, Bozovich G, Daroca A, et al. Randomised trial of roxithromycin in non-Q-wave coronary syndromes: ROXIS pilot study. ROXIS Study Group. Lancet. 1997; 350:404-7.

Ischemic heart disease is the number 1 killer worldwide (Table 1): It accounted for 1 in 8 of the 50 million deaths reported globally in 1990. Many other causes of common fatal diseases are clearly infectious, and growing evidence indicates that coronary artery disease also may be infectious. If so, even coronary disease might be amenable to antibiotic therapy, in a manner analogous to the precedent established by Helicobacter pylori and peptic ulcer disease.

Early studies provided serologic evidence of associations between high titers of antibody to Chlamydia pneumoniae and both coronary artery disease and cerebrovascular disease [12, 13]. However, serologic tests for C. pneumoniae are not particularly accurate, although the observations have been consistent.

Subsequently, C. pneumoniae was found in atherosclerotic plaques by a variety of nonserologic techniques [14]. The infection rate ranged from 35% to 86% in various reports compared with 0% in the arterial walls of controls. Finally, C. pneumoniae was recovered in culture from atheromatous plaque.

If C. pneumoniae infection contributes to the pathogenesis of ischemic heart disease, then treatment against this organism should improve outcomes; the purpose of this study was to test that hypothesis. In a multicenter study, 202 patients with unstable angina or non-Q-wave myocardial infarction were randomly assigned to receive either the macrolide antibiotic roxithromycin (150 mg twice daily orally) or placebo for 30 days. Patients were followed for 1 month to assess the primary end points of cardiac ischemic death, myocardial infarction, or severe recurrent ischemia.

At the end of the study, recurrent ischemia had occurred in 5.4% of those receiving placebo and 1.1% of those receiving roxithromycin. Myocardial infarction occurred in 2.2% and ischemic death in 2.2% of those receiving placebo compared with none of those receiving the antibiotic. Because of the small numbers of study participants, the statistical power of this pilot study was weak. However, the clinical implications are provocative, to say the least.

Of course, it would be premature to start using antibiotics in patients with atherosclerosis on the basis of such slender evidence. Other investigators have also implicated cytomegalovirus or herpes simplex virus in coronary disease. Numerous diagnostic and therapeutic studies of an infectious cause of coronary disease are under way, and we can expect much more information in the next year or two.

Human Herpesvirus-8 Seemed To Be Sexually Transmitted

Martin JN, Ganem DE, Osmond DH, et al. Sexual transmission and the natural history of human herpesvirus 8 infection. N Engl J Med. 1998; 338:948-54.

Human herpesvirus-8 (HHV-8) was originally reported as a possible cause of Kaposi sarcoma because its DNA sequences were found in 90% of patients with AIDS-associated Kaposi sarcoma [15]. Further studies found HHV-8 in several other types of sarcoma, including the Kaposi sarcoma occurring in transplant recipients and that endemic to Africa and the Mediterranean region. The virus was also implicated in body cavity lymphomas.

Still, little is known about the prevalence of HHV-8 in the general population, its modes of transmission, or its natural history. Martin and colleagues sought to answer those questions. The source of their study was the San Francisco Men's Health Study, which began in 1984 and has been a major source of epidemiologic information about patients with HIV infection. Researchers examined the baseline serum specimens that were stored from 400 patients with and 400 patients without HIV infection. They measured antibody levels in these serum samples and then followed the patients' course.

Antibody to HHV-8 was found in 38% of the homosexual men and none of the heterosexual men. Other associations with HHV-8 infection are shown in Table 2. The strongest associations indicate that HHV-8 infection is primarily a sexually transmitted disease.

The data from clinical follow-up were perhaps even more interesting. Among all patients who were eventually HIV positive, 30% developed Kaposi sarcoma and all had HHV-8 antibody in their baseline samples. Overall, for those who were positive for both HIV and HHV-8 at the study's inception, the 10-year risk for developing Kaposi sarcoma was 50%.

I believe that HHV-8 will prove to be a clinically diverse virus, much like Epstein-Barr virus. It is almost certain to cause Kaposi sarcoma, and it is associated with body cavity lymphomas. It may also play a role in multiple myeloma and sarcoidosis [16]. These latter associations may or may not hold up over time, but HHV-8 is likely to receive much clinical attention in the future.

Biological Warfare Is a Medical Issue

Franz DR, Jahrling PB, Friedlander AM, et al. Clinical recognition and management of patients exposed to biological warfare agents. JAMA. 1997; 278:399-411.

Most clinicians probably think little about biological warfare in their day-to-day practices. But the issue is one of growing concern, and the physician's examination room may well be the site of the first clues to an outbreak of such violence.

The entire 6 August 1997 edition of the Journal of the American Medical Association was devoted to biological warfare. This review by Franz and colleagues, published in that issue, pointed out that because perpetrators are unlikely to announce their actions and because most biological agents have incubation periods, physicians whose patients present with vague symptoms will probably be the first to recognize that these agents have been used.

At least 10 agents have been categorized as possible biological weapons, but most experts on biological warfare agree that the 2 or 3 organisms most likely to be used in an attack are anthrax, smallpox, and perhaps plague or tularemia.

A summary of the use of anthrax demonstrates the malevolently complex nature of biological warfare. Weaponry containing anthrax has been developed or stored by the United States, Japan, the United Kingdom, and Russia. An accident at the biological warfare facility in the Soviet Union in 1979 resulted in 79 cases of anthrax and 66 deaths. In 1991, Iraq admitted to having 8000 L of anthrax spores at a concentration of 10⁹ spores/mL, enough to kill the entire population of the world about 1000 times. The Japanese cult that sprayed subways with a chemical agent in 1995 also had stores of anthrax. Although the cult's leader remains in prison, the cult still has at least 2000 members.

Inhalation of anthrax spores produces respiratory distress and fever after 1 to 5 days. About 50% of affected persons develop meningitis, and about 80% die. Chest radiography shows no infiltrate but often shows mediastinal widening. The organism is not found in sputum. It can be found in blood but may take several days to grow. The treatment is penicillin, doxycycline, or ciprofloxacin, but it must be given early and even then is often ineffective. An anthrax vaccine exists, but supplies are limited.

What would happen if anthrax were used as a military weapon? We would first need to recognize the clinical cases and initiate therapy as rapidly as possible. In the exposed but non-ill population, we would need to vaccinate everyone and administer an antibiotic for 30 days until antibody titers reached adequate levels. This scenario underscores the reasons why biological warfare experts are so fearful: Anthrax infection is difficult to diagnose, treatment of infected patients is often unsuccessful, vaccine stores are insufficient, and no national emergency program exists to guide an organized response.

The challenge with smallpox would be far greater because of the extraordinary efficiency of person-to-person transmission of this virus. Other important features of an attack involving smallpox are the almost complete vulnerability of the population to infection, the high mortality rate associated with infection, the lack of effective therapy, and our sparse stores of vaccine. Again, physicians would be the first responders, and most physicians are unfamiliar with the clinical features of the disease. The salient features of the three major infectious agents are briefly summarized in Table 3.

HIV Disease

Clinical advances against HIV continue at a rapid pace. Guidelines are provided by the U.S. Department of Health and Human Services, with updates published at 1- to 2-month intervals. It is almost impossible to remain truly up-to-date with HIV care unless one studies the field full time, but several summary points are worth particular emphasis.

Most important, of course, is the status of the current drugs available to treat HIV infection. At this writing, 11 drugs are available-5 nucleoside reverse transcriptase inhibitors (NRTIs), 4 protease inhibitors, and 2 non-nucleoside reverse transcriptase inhibitors (NNRTIs). By the end of 1998, 2 or 3 more drugs will probably be added to this list (Table 4).

Protease Inhibitor Trial Marked End of a Research Era

Hammer SM, Squires KE, Hughes MD, et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. AIDS Clinical Trials Group 320 Study Team. N Engl J Med. 1997; 337:725-33.

A pivotal study in 1997 was reported by the AIDS Clinical Trials Group 320 Study Team. These investigators enrolled 1156 patients with CD4 cell counts of 200 cells/mm³ and randomly assigned them to receive either two nucleoside analogues alone (zidovudine, 600 mg/d, plus lamivudine, 300 mg/d) or the two nucleoside analogues plus the protease inhibitor indinavir (2400 mg/d). Stavudine was allowed as a substitute for zidovudine. Primary outcomes were development of AIDS or death.

After a median of 38 weeks, 11% of those receiving only the two nucleoside analogues and 6% of those receiving the two nucleoside analogues plus the protease inhibitor had one of the outcomes (relative risk reduction, 50% [95% CI, 24% to 67%]). The responses of CD4 cell counts and plasma HIV-1 RNA levels paralleled the clinical results.

This study was pivotal for two reasons. First, it showed that three drugs-two nucleoside analogues and a protease inhibitor-are superior to two nucleoside analogues alone. This was not surprising. Second, and perhaps more important, this is probably the last study that will use clinical outcomes, such as progression to AIDS or death, as end points in an antiretroviral clinical trial. At the completion of this study, the U.S. Food and Drug Administration ruled that all future treatment studies simply need to show a virologic response. An analysis of 22 studies involving more than 5000 patients showed that the surrogate measure of virologic response matched clinical response very closely.

Treatment Guidelines Paralleled One Another

Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. Department of Health and Human Services and Henry J. Kaiser Family Foundation. MMWR Morb Mortal Wkly Rep. 1998; 47(RR-5):43-82.

Guidelines for the use of antiretroviral agents in HIV-infected adults and adolescents. Department of Health and Human Services and the Henry J. Kaiser Foundation. Ann Intern Med. 1998; 128(12 pt 2):1079-100.

Numerous HIV disease guidelines have been published in 1998. They address the use of antiretroviral agents in pediatrics [17] and in pregnancy [18], recommendations for chemoprophylaxis after occupational exposure [19], and the use of antiretroviral agents in adults, which is cited above.

These DHHS guidelines are created by a 35-member consensus panel. They are updated monthly by a steering committee and then voted on by the entire panel. Table 5 summarizes the contents of the current DHHS guidelines and compares them with the International AIDS Society Guidelines published in 1998 [20]. The two panels provide markedly similar recommendations about when to begin therapy, what treatments to use first, and when to consider changing therapy. All patients with advanced or symptomatic HIV disease should receive aggressive antiretroviral therapy. For patients with early HIV disease, the issues are more complex. However, there is general consensus on the need to be more aggressive and to strive for complete viral suppression. Many of us think that recognition of the acute HIV syndrome is especially important because it may provide a unique opportunity for successful early aggressive therapy, and growing evidence suggests that this approach may be wise [21].

The goal of therapy is to eliminate the virus or at least bring it down to undetectable levels. The best predictor for a prolonged response is the nadir of the viral burden. In practice, reaching undetectable levels of virus is often impossible, but the lower the level, the better the outcome.

Two other issues were addressed in this guideline. For pregnant women, the old policy was to give zidovudine to prevent vertical transmission; this therapy was initiated after the first trimester. The new policy is to treat the mother for HIV infection according to the previous adult guidelines and to try to include zidovudine because it is known to be effective in preventing vertical transmission [18]. For health care workers who have needlestick exposures, the old policy was to use zidovudine plus lamivudine with or without indinavir. The new policy will be to stratify injuries according to level of risk for infection. The options will range from no postexposure prophylaxis to the use of triple therapy with protease inhibitors [19].

Analysis of Outcomes Demonstrated Benefit of Newer Drugs

Palella FJ Jr, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med. 1998; 338:853-60.

The new therapies have had a dramatic effect. Mortality rates from HIV disease in the United States decreased 12% in 1996 and 44% in the first half of 1997. Perinatal transmission is down 43%.

In this nine-center study, 1255 patients with CD4 cell counts of less than 100 cells/mm³ received standard-of-care antiretroviral therapy and prophylaxis against opportunistic infection. The study began early in 1994 and concluded in mid-1997; this period saw rapid clinical advances.

Mortality decreased from 29.4 per 100 person-years in 1995 to 8.8 per 100 person-years by mid-1997, a 70% decrease. The incidence of any of the three major opportunistic infections-Pneumocystis carinii pneumonia, Mycobacterium avium complex disease, and cytomegalovirus retinitis-decreased by 83%. The findings were consistent among all groups, except that mortality among those with private insurance was lower than that among persons covered by Medicaid.

At our institution, we have seen similar data. Furthermore, despite the high cost of these newer therapies, the cost of care has decreased along with morbidity and mortality, mostly because of decreased rates of hospitalizations.

Finally, for those interested in further advances in infectious diseases, our division actively maintains two frequently updated World Wide Web sites, one dedicated to general infectious diseases (http://www.hopkins-id.edu) and one devoted to HIV disease (http://www.hopkins-aids.edu).

Dr. Roberts (Series Editor): Madrona Medical Group, 3199 Steller Court, Bellingham, WA 98226.