Correspondence.

(Letters to the Editor are welcomed. They may report new clinical or laboratory observations and new developments in medical care or may contain comments on recent contents of the Journal. They will be published, if found suitable, as space permits. Like other material submitted for publication, letters must be typewritten, double-spaced, and submitted in duplicate. They must not exceed two typewritten pages in length. No more than five references and one figure or table may be used. See "Information for Authors" for format of references, tables, and figures. Editing, possible abridgment, and acceptance remain the prerogative of the Editors.)

Successful Treatment of Cryptococcal Osteamyelitis and Paraspinous Abscess

To the Editor: I enjoyed the case history, "Successful treatment of cryptococcal osteomyelitis and paraspinous abscess with fluconazole and flucystosine." (1) I do take exception, however, to the implication that the patient has been completely cured of this concern with the medical treatment alone. I do agree that the active infection has probably been successfully treated without surgical intervention, which is noteworthy. The resultant cervicothoracic kyphotic spine deformity (as seen in the magnetic resonance imaging of Figure 2) is too severe to be considered stable, however. I think (and suspect most spine surgeons would agree) that this patient is going to need stabilization of her cervicothoracic junction. This is particularly true in such a young patient, unless there are other medical complexities that would preclude her undergoing surgery. I agree that the surgery is going to be a significant undertaking, but I do think that it is necessary. If nothing is done, this patient will most likely have pr ogressive kyphosis that will lead to unacceptable chin-on-chest posture and/or neurologic impairment.

Lloyd I. Maliner, MD

Neurology and Neurosurgery Associates, PA

50 2nd St SE

Winter Haven, FL 33880

Reference

(1.) Cook PP: Successful treatment of cryptococcal osteomyelitis and paraspinous abscess with fluconazole and flucytosine. South Med J 2001; 94:936-938

Medicine and the Sin Tax

To the Editor. I entered the field of medicine in the 1980s, when medical costs were becoming a real issue for the country. During my training in emergency departments, intensive care units, operating rooms, and on the hospital wards, I wondered what all the fuss over costs was about. I saw that not a single patient sent to the emergency department for an injury in a car wreck had been wearing a seatbelt. I saw that a huge proportion of them were driving while intoxicated or were hit by an intoxicated driver. I saw that every patient of mine with lung cancer, emphysema, or life-threatening heart disease smoked cigarettes. I saw that about half of all of my patients fell into 1 or more of those 3 categories (and this was before the effects of second-hand smoking were known). The solution to the growing cost problem, to me, was a no-brainer: taxes on cigarettes, alcohol, and beltless and reckless driving ("sin" taxes) to pay those costs. Eventually, because of the financial incentive, disease and injury would d ecline, thus reducing demand for health care. The supply of health care, and therefore the cost, would automatically follow. No problem.

Instead, health maintenance organizations try to keep the cost of health care low by reducing the supply of health care, rather than the demand. They do this, in part, by disallowing certain medications and treatments, assuming they know what is in the patient's best interest better than the physician does. How this is supposed to heal the sick is beyond me. Isn't it obvious, though, that the best way to reduce the cost of health care is to first reduce the demand for health care?

Fact: Tobacco is the single most preventable cause of death and illness in the country. Trauma is next, either from deliberate violence (for children or young adults) or from car crashes (for those in their 20s and older, often involving alcohol). Taxing tobacco, alcohol, and serious moving violations in amounts equal to their cost to society in medical terms, and then directing those funds solely to pay for the diseases and injuries they cause and nothing else (definitely not to the general treasury) would dramatically reduce health insurance companies' outlays, since they would no longer be obliged to pay for costs related to those factors. Premiums would fall proportionately, as directed by the appropriate government oversight agency. Nobody likes a new tax, but legislatively mandating that this tax money only be used to pay the costs their sources incur and nothing else is the only fair way to do it, and would nullify the 'just another tax" criticism. People would thus have a choice: they would pay an amo unt of tax dependent upon how much drinking, smoking, and bad driving they chose to do. This money would fund the consequent medical costs and nothing else. We know that the vast majority of adult tobacco users got hooked as children because cigarettes are cheap. The tax would make cigarettes too expensive for most kids, until they become adults with jobs and responsibilities, when many would see the folly of contributing $5 or $10 a day for decades toward their own pathetic and miserable doom. The cost of treating the rest of our many (nonpreventable) afflictions would come from health insurance.

Barry R. Bloom, dean of the Harvard School of Public Health, is a man who should know public health statistics. 1-le wisely submits that the only way to solve the health care expense crisis in America is to focus on not getting sick or injured in the first place. In a recent Newsweek article, he stated that half (half.!) of all deaths in the United States are preventable, citing tobacco as the cause of 19% and alcohol as the cause of 5% of all deaths.' If a quarter of all deaths are caused by common, recreational substance use, it is clear to me that taxing those substances to the extent that they generate expenses by causing disease, and applying those funds to pay for only those diseases, would have the following clearly beneficial effects: it would reduce the health insurance bill of the nation, allowing more people to afford private health insurance; reduce the ranks of the uninsured, as well as the government burden of Medicare and Medicaid; and ultimately improve the health of the nation by providing fi nancial incentive to avoid those destructive behaviors. Furthermore, the Internal Revenue Service would send this revenue right back to the people by paying the medical expenses so generated.

The practical matter of implementing such a policy would involve collecting the data that already exist on the relative contribution each of these behaviors makes to various diseases and injuries. Divide the number of these "behavior units" (eg, packs of cigarettes) in a year into their calculated annual medical cost. This would be the unit tax. Phase it in gradually, perhaps over a period of 5 years, to allow people time to break their habits. Gradual implementation would also allow those employed in the tobacco, alcohol, and trauma business (including physicians) to make the appropriate economic adjustments.

Is this a regressive tax, since proportionately more automobile accidents and abuse of alcohol and tobacco occur among people at lower income levels? Maybe, but who could legitimately call a concept that will reduce the burden of mortal suffering and disease a regressive one? This is no high-minded monologue. This proposal goes straight to individual responsibility for one's own choices. Moreover, as a medical practitioner, I am horrified and depressed at the medical decision-making now being carried out by the health maintenance organizations in an effort to spend less and impress their stockholders. As a consumer, I resent paying high health insurance premiums to cover the medical care made necessary by the habitually unthinking or careless choices of others. Don't you?

The last question left on sin taxes is, which politician has the guts to propose them?

Marshall Millman, MD, PhD

2005 Maplemere Dr

Nashville, TN 37215

Reference

(1.) Bloom BR: The wrong rights. Newsweek, Oct 11, 1999

Informed Consent for Circumcision

To the Editor: Binner et al (1) document that material prepared by the American Academy of Pediatrics in the brochure, Circumcision Information for Parents, (2) is inadequate to persuade parents to avoid circumcision. This is not surprising, because the information provided by the brochure does not satisfy the legal criteria for informed consent.

Binner et al (1) did not show that parents cannot be dissuaded from circumcision by sufficient information. Enzenauer et a1 (3) counseled parents of boys with videotaped presentations regarding circumcision. The videotaped counseling achieved a significant reduction in the number of parents who elected to have their sons circumcised. Thus, with sufficient information, a reduction in the high incidence of nontherapeutic circumcision indeed is possible and may be achieved.

Informed consent is governed by court decisions. Generally, the doctor must disclose all material information that the patient or his proxy representative would deem relevant about the benefits, risks, and possible outcomes of the proposed treatment and any alternative treatments. Medical practitioners who fail to provide sufficient information may be liable in negligence. (4)

The American Medical Association says that "major factors in parental decision-making are the father's circumcision status, opinions of family members and friends, a desire for conformity in their son's appearance, and the belief that the circumcised penis is easier to care for with respect to local hygiene." (5) Obviously, such decisions are not made in the child's best interest. The validity of the permission granted for circumcision based on such decisions may be challenged on the basis that they are made in the parents' interests rather than the child's best interest.

Medical societies are strengthening their advice to their fellows regarding informed consent for nontherapeutic neonatal circumcision. The American Academy of Family Physicians recently issued a new position statement on neonatal circumcision. The statement cautions doctors to discuss the benefits and risks with parents. (6) In Canada, the Saskatchewan College of Physicians and Surgeons recently warned its fellows that medical practitioners must tell parents that male neonatal nontherapeutic circumcision is not a recommended procedure. (7)

Thus, it is now clear that medical doctors must impart all material information to the parents, not only about circumcision, but also about the option not to circumcise. This information must be far more extensive than the scanty information contained in the American Academy of Pediatrics brochure. In the case of nontherapeutic circumcision of a boy, electing not to have circumcision performed is a reasonable, practicable, and viable alternative, so patients or their proxy representatives must be given all material information about circumcision and its risks, potential benefits, and possible outcomes; they must also be informed about the risks, potential benefits, and possible outcomes of remaining uncircumcised. No doctor should rely on such grossly inadequate material as a handout prepared by the American Academy of Pediatrics.

George Hill, BA

Executive Secretary

Doctors Opposing Circumcision

2442 NW Market St, Suite 42

Seattle, WA 98107

References

(1.) Binner SL, Mastrobattista JM, Day M, et al: Effect of parental education on decision-making about neonatal circumcision. South Med J 2002; 95:457-461

(2.) American Academy of Pediatrics: Circumcision Information for Parents. Elk Grove Village, Ill, American Academy of Pediatrics, 2001

(3.) Enzenauer RW, Powell JM, Wiswell TE, et al: Decreased circumcision rate with videotaped counseling. South Med J 1986; 79:717-720

(4.) Povenmire R: Do parents have the legal authority to consent to the surgical amputation of normal, healthy tissue from their infant children? the practice of circumcision in the United States. J Gender Soc Pol Law 1998; 7:87-123

(5.) Council on Scientific Affairs, American Medical Association: Report 10: Neonatal Circumcision. Chicago, American Medical Association, 1999

(6.) Commission on Clinical Policies and Research: Position Paper on Neonatal Circumcision. Leawood, Kan, American Academy of Family Physicians, 2002. Available at http://www.aafp.org/policy/camp/4.html. Accessed April 21, 2002

(7.) Kendel DA: Caution Against Routine Circumcision of Newborn Male Infants (Memorandum to physicians and surgeons of Saskatchewan). Saskatoon, College of Physicians and Surgeons of Saskatchewan, February 20, 2002

Treatment of Retroperitoneal Fibrosis With Tamoxifer

To the Editor We present a case of retroperitoneal fibrosis that was treated successfully with tamoxifen. A 35-year old woman was diagnosed 8 years ago with hypertension Intravenous pyelogram showed bilateral hydronephrosis involving the left side more than the right side. Renal perfusion scan revealed diminished function of both kidney with a glomerular filtration rate of 68 mL/min/[m.sup.2] Her initial laboratory results revealed a normal serum creatinine level of 0.9 mg/dL.

Results of an abdominal computed tomography (CT) scan confirmed bilateral hydronephrosis and a mass surrounding the distal part of the left ureter. An exploratory laparotomy with biopsy of the mass was performed, and results revealed retroperitoneal fibrosis. A J-stent was placed in the left ureter to relieve the obstruction.

Over a period of 6 years, the patient's condition deteriorated, with uncontrolled blood pressure and worsening kidney function. She underwent several laparotomies for lysis of adhesions around the ureters. In 1998, she moved to another area, and her new primary care physician suggested a trial of medical therapy. At that time, her physical examination was remarkable for blood pressure level of 158/95 mm Hg and a palpable mass in the left lower quadrant. Laboratory tests revealed a serum creatinine level of 1.9 mg/dL. She started tamoxifen therapy at a dose of 20 mg per day.

Repeated abdominal and pelvic CT scans showed progressive resolution of the mass. Over a period of 18 months, her blood pressure came under control, with improvement in her renal function to a glomerular filtration rate of 95 mL/min/[m.sup.2], and a decrease in her serum creatinine level to 0.9 mg/dL. Tamoxifen was subsequently tapered.

Diagnosing retroperitoneal fibrosis is difficult because the presenting symptoms and signs are usually nonspecific and physical findings are lacking. (1) The most common symptom is pain in the back, flank, and abdomen. Weight loss, weakness, claudication, and lower extremity swelling are less common. More than half of the patients have azotemia at presentation. (1)

A classic radiologic finding, seen in approximately two thirds of patients, is bilateral ureteral obstruction leading to hydronephrosis. Magnetic resonance imaging documents the infiltrating process in the retroperitoneum. (1) Intravenous pyelography can demonstrate medial deviation of one or both ureters at the L3 to L4 level. Biopsy is the only way to prove the diagnosis and rule out malignancy. Histologically, the presence of thick, hyalinized bands, a prominent inflammatory infiltrate, and relatively low cellularity distinguishes retroperitoneal fibrosis from other intra-abdominal pathologies.

Patients afflicted with idiopathic retroperitoneal fibrosis are difficult to manage surgically because of the associated morbidity. (2) In addition, these procedures are palliative and have no effect on the progression of the disease. While the medical management of retroperitoneal fibrosis is controversial because of the lack of randomized controlled studies, corticosteroids and immunosuppressive therapy have been tried.

The rationale for tamoxifen therapy stems from its use to alleviate ureteral obstruction secondary to metastatic breast cancer in the retroperitoneum. Although its role in the treatment of other proliferative disorders has been unclear, our report shows a significant benefit in using tamoxifen to treat a patient with retropentoneal fibrosis. In our case, as in other cases reported in the literature, the mass regressed in size after treatment with tamoxifen, and ureteral obstruction subsided.

There are different hypotheses which may explain the response of retroperitoneal fibrosis to tamoxifen. Tamoxifen increases the synthesis and secretion of transforming growth factor-[beta] (TGF-[beta]), an inhibitory growth factor, by human fetal fibroblasts in vitro, and increases the extracellular concentration of this growth-inhibiting peptide in human breast cancers in vivo. (3) In idiopathic retroperitoneal fibrosis, fibroblasts and immune cells present in the inflammatory mass may increase their secretion of TGF-[beta], which may then decrease the size of the fibrous mass. (3)

Tamoxifen also inhibits protein kinase C, which is an obligatory mediator of cell proliferation in mouse lymphoma cells. (4) Other possible mechanisms of action include inhibition of calmodulin, (4) blockage of a growth-promoting histamine-like receptor, or the reduction of epidermal growth-factor production. (5)

Due to the simplicity, effectiveness, and safety of tamoxifen therapy, it should be considered as a treatment for retroperitoneal fibrosis. The long-term effects are still uncertain, however, because few cases have been reported.

Jameela Al-Salman, MD

A. Rashid Makhdomi, MD

Department of Medicine

Easton Hospital

250 S 21st St

Easton, PA 18042

References

(1.) Jaffer A, Calabrese L: Severe back and abdominal pain in a 44-year-old woman. Cleve ClinJMed 1998; 65:515-518

(2.) Owens LV, Cance WG, Huth JF: Recroperitoneal fibrosis treated with tamoxifen. Am Sing 1995; 61:842-844

(3.) Spillane RM, Whitman GJ: Treatment of retroperitoneal fibrosis with tamoxifen. AJR Am J Roentgenol 1993; 164:515-516

(4.) Clark CP, Vanderpool D, Preskitt JT et al: The response of retroperitoneal fibrosis to tamoxifen. Surgery 1991; 109:502-506

(5.) Loffeld RJ, Van Weel TF: Tamoxifen for retroperitoneal fibrosis. Lanai 1993; 351:382

Fatal Granulomatous Bronchopneumonia Complicated by Acute Renal Failure 

To the Editor. Tularemia is a zoonosis caused by the small, facultative, gram-negative, intracellular coccobacilus, Francisella tularensis. Incidence in the United States is less than 1 per million; transmission is related to tick bites nd to hunting and skinning rabbits. Human-to-human transmission is rare, but possible. Tularemia produces several specific syndromes, including ulceroglandular fever, typhoidal tularemia and pneumonic tularemia. Acute mal failure (ARF) has rarely been described in relation to tularemia. The common causes of pulmonary-renal syndromes include Wegener's granulomatosis, Goodpasture'syndrome, and other vasculitides. Infectious agents such as malaria, Rocky Mountain spotted fever, typhoid, and leptospirosis can have similar presentations. (12)

A 46-year-old African American man was transferred to our intensive care unit with a rapidly progressive febrile illness, vomiting, and diarrhea. Before transfer, he was found to have pneumonia, gastrointestinal bleeding, elevated transaminase levels, and progressive azotemia. On admission to our unit, the patient was drowsy. His temperature was 96.4[degrees] F, his heart rate was 121 beats/min, he was tachypneic, and there was blood in his nasogastric tube. Chest examination revealed bilateral inspiratory crackles posteriorly. The abdomen was distended and tender, and his rectal tube contained bloody fecal material. A Foley catheter contained scanty, blood-tinged urine. Results of tests for urinary legionnaire's-disease antigen and Lyme-disease antibody were negative at admission. Serum protein electrophoresis results were normal and culture results were negative. Results of other tests, including tularemia titers, psittacosis titers, Hantavirus serology, antinuclear antibodies test, human immunodeficiency v irus serology, and perinuclear antineutrophilic cytoplasmic antibody (p-ANCA), cytoplasmic antineutrophilic cytoplasmic antibody (c-ANCA) tests, rapid plasmin reagin, erythrocyte sedimentation rate, C3, C4, and urine electrolyte levels, were pending. On admission, pertinent laboratory data included the following values: white blood cell (WBC) count, 17.5 x [10.sup.3]/[micro]L (89% polymorphs); hematocrit, 24.2%; total bilirubin, 9.6 mg/dL; albumin, 1.9 g/dL; phosphorus, 10.2 mg/dL; chloride, 113 mEq/L; carbon dioxide, 11 mmol/L; blood urea nitrogen, 139 mg/dL; creatinine, 9.4 mg/dL; aspartate aminotransferase, 1,384 U/L; alanine aminotransferase, 610 U/L; lactic acid, 12.7 mg/dL; amylase, 209 U/L; lipase, 342 U/L; creatinine phosphokinase, 229 IU/L; myoglobin, 1,872 ng/mL. Arterial blood gases were pH, 7.20; p[CO.sub.2], 24 mm Hg; p[O.sub.2], 139 mm Hg; bicarbonate level, 9 mEq/L; and oxygen saturation, 98.6% on supplemental oxygen. Results of coagulation tests were normal. Chest radiograph showed extensive, multifocal, peripheral-airspace opacities, consistent with necrotizing pneumonitis. Computed tomography (CT) scan of the chest demonstrated multifocal, wedge-shaped regions of consolidation and coalescent pulmonary nodules with cavitation.

Respiratory isolation was initiated. Rectal bleeding continued; abdominal CT scan showed bowel-wall thickening, but no perforation. The patient's respiratory distress worsened, and it was necessary to intubate him. Blood transfusion was performed, and daily hemodialysis was initiated. Intravenous treatment with doxycycline and levofloxacin continued. Examination of cerebrospinal fluid revealed protein level of 88 mg/dL, normal glucose level, WBC count of 6/[micro]L (all polymorphonuclear neutrophils), and 8,550 red blood cells/[micro]L (traumatic tap). Endoscopy demonstrated multiple bleeding ulcers in the stomach and esophagus. Urine output was 66 [cm.sup.3] during the first 12 hours. By day 2, prothrombin time and international normalized ratio (INR) were prolonged, requiring infusions of fresh frozen plasma (FFP) and cryoprecipitate. By day 3, with all culture results negative, a diagnosis of Wegener's granulomatosis was strongly suspected. Intravenous methylprednisone was started empirically. The next day , p-ANCA and c-ANCA test results were negative. Laboratory evaluation showed features of disseminated intravascular coagulation. Following infusions of packed red blood cells and FEP, a renal biopsy was done. Pathologic examination revealed acute tubular necrosis (ATN), but no glomerulonephritis (Figure). The next day, the Mayo Clinic Laboratories reported growth of Francisella tularensis from earlier bronchoscopy washings. Soon it was discovered that the patient was a hunter and had skinned deer for other hunters. Treatment with intravenous ciprofloxacin, bactericidal against Francisella tularensis, was begun. The drug of choice for tularemia, streptomycin, was avoided because of concerns about ARF. (2) The patient continued to deteriorate. Intravenous gentamicin was added to his medications. His pupils became nonreactive; a head CT confirmed brainstem herniation, and he died 11 days after admission. All culture results and other outstanding blood test results were negative. Hepatitis serology showed only he patitis B surface antibody. CSF studies for cryptococcal antigen, herpes polymerase chain reaction, and Venereal Disease Research Laboratory test for syphilis were negative. Result of an unpaired serology test for Francisella tularensis was 1:20. Autopsy revealed extensive bronchopneumonia, hemorrhagic stomach, and necrotic, hemorrhagic spleen with depletion of the lymphoid lineage. The liver was passively congested, and both kidneys were hemorrhagic and showed marked acute tubular necrosis (ATN).

Possible diagnoses in our patient included infective and noninfective etiologies for acute pulmonary-renal syndrome. A rapid enzyme-linked-immunosorbent-assay-based screening for anti-glomerular- basement-membrane antibody and ANCA tests have proven useful in making this differentiation. (3) As expected with the negative ANCA results, there was no glomerulonephritis. Pulmonary-renal disorders can be extraordinarily fulminant, leading to irreversible organ failure and death in a matter of hours to days. This calls for an aggressive approach to very early diagnosis. Severe cases of pulmonary-renal syndromes, frequently complicated by a bleeding diathesis, make open biopsies very risky procedures. Tilley et al (2) diagnosed tularemia in a former painter with a history of exposure to a dead rabbit with nonoliguric ARF and pneumonia using rising antibody titers. We became aware of our patient's exposure history only after the tularemia-culture report was available. Occupational history, travel history, and any oth er unusual exposures are important in the evaluation of such patients.

The pathology of ARF in tularemia is not well known. A recent review of the infectious diseases associated with renal failure failed to include tularemia. (1) In his review of 600 patients with tularemia, Foshay (4) described only two patients with significant renal abnormalities. Pathologic studies have described ATN, exudative glomerulonephritis, and interstitial nephritis in tularemia; rhabdomyolysis has rarely been implicated. Our patient did not have rhabdomyolysis; he had rapidly progressive ARF with proteinuria and hematuria secondary to ATN.

In pulmonary-renal syndromes, infectious causes should receive as much emphasis as immunologic causes. This could call for early isolation, appropriate cultures, and paired serology testing. Serology assays may be the only diagnostic test results in such cases. (2) The latter is particularly true in tularemia, where positive blood culture results remain extremely rare. (5) Once tularemia is suspected, appropriate broad-spectrum antibacterial therapy should be initiated.

Macaulay Onuigbo, MD, MSc

Michael Hise, MD

Emilio Ramos, MD

Division of Nephrology

Nelson Traong, MD

Pamela Amelung, MD

Department of Critical and Pulmonary Care

Cinthia Drachenberg, MD

Division of Pathology

University of Maryland School of Medicine

22 S Greene St

Room N3W143

Baltimore, MD 21201-1595

References

(1.) Eknoyan C, Dillman RO: Renal complications of infectious diseases. Med Clin North Am 1978; 52:979-1003

(2.) Tilley WS, Garman RW, Stone WJ: Tularemia complicated by acute renal failure. South Med J 1983; 76:273-274

(3.) Westman KW, Bygren PG, Eilert I, et al: Rapid screening assay for anti-GBM antibody and ANCAs; an important tool for the differential diagnosis of pulmonary syndromes. Nephrol Dial Transplant 1997; 12:1863-1868

(4.) Foshay L: Tularemia: a summary of certain aspects of the disease including methods for early diagnosis and the results of serum treatment in 600 patients. Medicine 1940; 19:1-83

(5.) Proveoza JM, Klotz SA, Penn RL: Isolation of Francisella tularensis from blood. J Clin Microbiol 1986; 24:453-455

Relapsing Group B Streptococcal Bacteremia

To the Editor: We report a case of a 61-year-old woman with end-stage liver disease due to hepatitis C who presented with a 1-day history of right lower quadrant pain and subjective fever. She had experienced 2 previous episodes of group B streptococcal bacteremia. The first episode occurred 10 months previously, when the patient presented with nausea and chills. Two sets of blood cultures were positive for group B streptococci (GBS). She was treated empirically with intravenous piperacillin and tazobactam until sensitivities were determined. Computed tomography (CT) of the abdomen and pelvis, as well as a transesophageal echocardiogram, showed no abnormalities. The patient responded to therapy, and was treated with intravenous antibiotics for 1 week. Subsequent blood cultures were negative for GBS, and the patient was treated with cephalexin for 2 more weeks. The second episode occurred 2 months before admission, when the patient presented with a 1-day history of nausea, vomiting, and headache. She was empir ically treated with intravenous vancomycin and ceftriaxone sodium. Blood cultures were again positive for GBS. Findings on CT scans of the abdomen and head were unremarkable. A transthoracic echocardiogram revealed no vegetations. The patient's condition improved, and her medication was switched to intravenous penicillin G on the second hospital day. The patient had pruritis within 24 hours of the medication change, so treatment was changed from penicillin G to vancomycin, and she continued therapy for 4 weeks.

The patient's only past surgery was a total abdominal hysterectomy and salpingo-oophorectomy 3 years previously that had been complicated by the need for a partial colectomy. She denied any history of intravenous drug use or blood transfusion.

On presentation, her temperature was 101.2[degrees]F, and her vital signs were stable. Physical examination was remarkable for icterus and a II/VI systolic ejection murmur at the left sternal border without radiation. Examination of the abdomen showed right lower quadrant tenderness with guarding. Pelvic examination revealed a residual cervical rim without exudate or motion tenderness. Results of laboratory tests revealed the following values: white blood cell count, 3,500/m[m.sup.3] hemoglobin, 12.3 g/dL; platelet count, 26,000/m[m.sup.3] serum glutamate oxaloacetate transaminase (SGOT), 74 U/I; serum glutamate pyruvate transaminase (SGPT), 32 U/I; albumin, 2.7 g/dL; glucose, 131 mg/dL; international normalized ratio (INR), 1.7.

Empiric therapy with intravenous vancomycin, aztreonam, and metronidazole was begun. Two sets of blood cultures were positive for GBS. The urine culture was also positive for > [10.sup.5] GBS. The cervical culture revealed rare growth of GBS, and fecal culture results were negative. Susceptibility testing revealed the organism to be pansensitive. After 48 hours, the patient defervesced, and the right lower quadrant pain resolved. No growth was observed in any of the blood cultures 3 days after admission, Findings on CT of the abdomen and pelvis were normal. A transthoracic echocardiogram revealed no vegetations, and an iridium scan revealed no focus of infection. Given the patient's questionable history of penicillin allergy, she had skin testing; the results were negative. She subsequently received 2 oral challenges of penicillin without reaction. She was treated with a 2-week course of intravenous penicillin G, followed by twice-daily prophylaxis with oral penicillin.

Group B Streptococcus is increasingly recognized as an emerging pathogen in nonpregnant adults; at least 50% of all GBS disease is seen in this population. As shown in our case, primary bacteremia is the single most common presentation, accounting for one third of cases. (1,2) Other sites of infection include the urogenital tract, skin, and lungs. The majority of cases of bacteremia are in older patients (over age 50). Older patients with GBS bacteremia have a high mortality rate, ranging from 30% to 67%. (1)

Approximately 4% of nonpregnant adults who survive an episode of GBS bacteremia and have at least 1 year of follow-up will have a second episode of the disease. (3) In patients with relapsing infections, the mean interval between episodes of bacteremia was 24 weeks; however, the intervals averaged 13 weeks if the bacteremia was caused by the same strain, versus 43 weeks if caused by a different strain. (3) Patients with primary bacteremia at initial presentation may have focal infections, such as endocarditis or osteomyelitis, discovered at their second presentation.

Patients who relapse with the same strain of GBS may have persistent gastrointestinal or genitourinary carriage, despite treatment. Asymptomatic genital and gastrointestinal colonization have been observed in both pregnant and nonpregnant women. Another possibility is that relapsing patients may have reacquired matching strains from a household contact. Relapsing patients with different strains are more likely to have host immune defects, predisposing them to recurrent infections. In one series of relapsing invasive GBS infections, the most common underlying medical conditions included cancer, diabetes, cirrhosis, and renal transplantation. (3)

In our case, the interval between the first and second episode of bacteremia was long (10 months), suggesting relapse with a different strain. This initial relapse may have been due to impaired host immunity due to cirrhosis caused by hepatitis C infection; however, the interval between the second and third episode of CBS bacteremia was only 2 months, suggesting relapse with a matching strain. During her third episode of bacteremia, the patient's cervical culture was positive for CBS 4 weeks after completion of intravenous vancomycin therapy. Although molecular subtyping could have been used to determine the CBS strains, it was not done in this case. (3)

The classical complement pathway and heat-stable opsonins are required for maximal opsonic activity against the major serotypes of CBS. (4) The low levels of complement associated with aging may play a role in the increased susceptibility to GBS disease in the older population. Deficient bacterial and opsonic activity, in correlation with low complement, has been recognized as a possible explanation for increased susceptibility to infections in patients with cirrhosis. (5) Although these deficiencies may increase the risk of subacute bacterial peritonitis, this is an unlikely source of recurrent bacteremia in our patient, given the absence of ascites on CT scans.

Episodes of primary bacteremia may be appropriately treated with a 10-day course of intravenous penicillin G (vancomycin can be used as an alternative agent). Patients with recurrent CBS bacteremia, however, should be routinely evaluated for focal infections, such as endocarditis or osteomyelitis, which require 4 to 6 weeks of antiobiotic therapy. Patients with recurrent bacteremia and without an identifiable source of infection should be considered for eradication of possible CBS carriage; however, the most efficacious antibiotic regimen and its effectiveness are unknown.

Michael T. Flannery, MD

Lara Winters, BS

University of South Florida College of Medicine 

Department of Internal Medicine

4 Columbia Dr

Suite 630

Harbourside Medical Tower

Tampa, FL 33606

References

(1.) Colford J, Mohle-Boetani J, Vosti K, et al: Group B streptococcal bacteremia in adults: five years' experience and a review of the literature. Medicine (Baltimore) 1995; 74:176-190

(2.) Munoz P, Llancaqueo A, Rodriquez-Creixems M, et al: Group B streptococcus bacteremia in nonpregnant adults. Arch Intern Med 1997; 157:213-216

(3.) Harrison L, Ali A, Dwyer D, et al: Relapsing invasive group B streptococcal infection in adults. Ann Intern Med 1995; 123:421-427

(4.) Shigeoka A, Hall R, Hemming V, et al: Role of antibody and complement in opsonization of group B streptococci. Infect Immun 1978; 21:34-40

(5.) Akalin H, Laleli Y, Telatar H: Serum bactericidal and opsonic activities in patients with non-alcoholic cirrhosis. Q J Med 1985; 56:43 1437