The acute renal failure required treatment with daily hemodial-
ysis, which subsequently was substituted with continuous ven-
ovenous hemofiltration because of hemodynamic deterioration.
Hemolytic uremic syndrome and thrombotic microangiopathy
were recognized since the acute renal failure was associated with
intravascular hemolysis, thrombocytopenia, and the presence of
6% schistocytes on the complete blood count, which decreased to
2% after plasmapheresis and administration of fresh frozen plas-
ma. Seventeen days after the onset of these symptoms, the patient
developed circulatory shock with lactic acidosis that failed to re-
spond to the administration of vasoactive agents. In addition, he
developed a persistent high-grade fever despite broad-spectrum
antibiotic therapy (tazocillin, ofloxacin, vancomycin, ampho-
tericin B). The patient died of multiorgan dysfunction syndrome
(respiratory, renal, hepatic) 17 days after the first symptoms ap-
peared. No autopsy was performed.
ommended. However, in severe cases, this therapy is often
unsuccessful, and patients eventually die (Table 1).4,5 Rawls
et al.5 reported five cases of sepsis following intravesical
BCG that progressed to multiorgan failure and death in
three patients despite administration of isoniazid, rifampin,
and streptomycin. Postmortem examination revealed BCG
organisms in the lung in only one patient (case 2). The two
latter patients who had equally progressive sepsis survived
with the use of cycloserine. Moreover, the bacteriologic
cultures are often negative, perhaps due to the antitubercu-
lar treatment. However, in some patients, biopsy shows
granuloma formations, which appear to be a hypersensitiv-
ity reaction since viable mycobacteria have not been re-
covered from these lesions. In these cases, treatment with
steroids may be more advisable, but it is not easy to distin-
guish a priori a hypersensitivity reaction from a true my-
cobacterial infection. Thus, Baba et al.8 reported a case of
multiorgan failure, disseminated intravascular coagulation,
and respiratory failure after administration of intravesical
BCG. This is similar to our case, but the patient recovered
with the administration of only antibiotics, not antitubercu-
lous therapy or corticosteroids.
The pathogenic mechanisms of the systemic reactions
related to BCG therapy are complex: they involve either a
systemic dissemination of the BCG-attenuated strain or an
immune disorder (hypersensitivity reaction) with an im-
portant release of proinflammatory cytokines (interleukins,
interferon, tumor necrosis factor, granulocyte-macrophage
colony-stimulating factor). Sepsis results from the sys-
temic absorption of BCG through the blood vessels of the
inflamed or disrupted urothelium. For this reason, BCG
should not be administered in the presence of cystitis or af-
ter a traumatic catheterization of the urinary tract. In our
patient, BCG was not administered after a transurethral re-
section or cystoscopic examination, procedures that can
play a role in the systemic dissemination of the bacillus (as
in the case reported by Baba et al.8). However, after the
seventh instillation, our patient developed a fever associat-
ed with chills, sweating, headache, and vomiting, which
resolved spontaneously. According to the recommenda-
tions reported in the literature,2 the BCG therapy should
have been discontinued at this point. In some of the cases
of BCG sepsis summarized in Table 1, patients experi-
enced an episode of fever with previous instillations of
BCG. The recommendations concerning BCG withdrawal
are not always respected, and accurate guidelines must be
provided to better inform physicians.
According to the Naranjo probability scale,10 this adverse drug
reaction was rated possible.
Discussion
Intravesical BCG immunotherapy is usually well toler-
ated, although 95% of patients experience minor adverse
effects, usually a flu-like syndrome, which appears after
the third injection and resolves spontaneously within 24
hours without specific treatment.3 Serious complications
have occasionally been reported, including hepatitis, renal
abscesses, retroperitoneal lymphadenopathy, miliary pul-
monary disease, and septic shock, all of which can be con-
sidered a systemic dissemination of BCG.4 BCG therapy
complications have been well studied in a large cohort of
2602 patients.2 In this study, sepsis occurred in 0.4% of
treated patients. The complication rate observed does not
correlate with the strain of BCG used. The authors speci-
fied that complications of BCG vaccination in cancer ther-
apy are relatively rare but are more frequent than in tuber-
culosis vaccination programs. Thus, while only one death
per 50 million patients occurred when BCG was given for
tuberculosis prophylaxis, a mortality rate of one death per
12 500 patients was observed when BCG was adminis-
tered in the treatment of cancer patients.
Recommendations of these authors2 concerning the
treatment of complications are: (1) fever <38.5 ˚C: no
treatment; (2) fever >38.5 ˚C for 12–24 hours and symp-
toms of allergic reactions (skin rash, arthralgia, migratory
arthritis): isoniazid 300 mg/d for three months; (3) acute,
severe visceral manifestations (granulomatous pneumoni-
tis, hepatitis): isoniazid 300 mg/d, rifampin 600 mg/d, and
ethambutol 1200 mg/d for six months; (4) severe sepsis
complicated by circulatory shock, acute respiratory distress
syndrome, and/or disseminated intravascular coagulopa-
thy: isoniazid 300 mg/d, rifampin 600 mg/d, ethambutol
1200 mg/d plus cycloserine 500 mg twice a day, and an in-
travenous bolus of prednisolone 40 mg.
Specific disseminated sepsis is known to be the most se-
vere complication following intravesical BCG instillation.
In such cases, patients usually present with symptoms typi-
cal of sepsis, including circulatory shock, acute respiratory
distress syndrome, and disseminated intravascular coagu-
lopathy. Since the strain administered is a BCG-attenuated
strain, triple or quadruple antituberculosis therapy is rec-
All cultures, especially the blood cultures, performed a
few days after the patient was admitted to the hospital re-
mained negative for mycobacteria two or three months lat-
er. In the course of the disease, the only positive culture
was a pulmonary culture positive for Candida tropicalis,
which was treated with amphotericin B. However, the first
episode of fever after the seventh instillation, associated
with the rapid onset of severe illness after the eighth instil-
lation, made us believe that BCG therapy, rather than a
fungal infection, was the cause of the complications devel-
■
■
1280
The Annals of Pharmacotherapy
2000 November, Volume 34