Nosocomial Outbreak of Burkholderia pickettii Infection Due to a Manufactured Intravenous Product Used in Three Hospitals
Jose Alberto Mariano
Received 23 May 1995;
revised 25 January 1996. This work was presented in part at the 3rd International Conference of the Hospital Infection Society held on 4-8 September 1994 in London and at the 34th Interscience Conference on Antimicrobial Agents and Chemotherapy held on 4-7 October 1994 in Orlando,
San Carlos University Hospital
, Profesor Martin Lagos sin,
. Clinical Infectious Diseases 1996
22:1092-5 1996 by The University of Chicago. All rights reserved. 1058--4838/96/2206-0031$02.00
From the Preventive Medicine and Clinical Microbiology Services, Hospital Universitario San Carlos; the Microbiology and Infectious Diseases Services, Hospital Severo Ochoa; and the Epidemiology Service
, Comunidad de Madrid,
Forty-six cases of nosocomial infection caused by Burkholderia pickettii were reported between June and November 1993 in three metropolitan hospitals in Madrid. A case-control study of the outbreak was conducted to identify its cause. Seventy-four percent of the patients were males; the mean age SD of the patients was 54 20 years. Sixty-five percent of the patients presented with some gastrointestinal disorder, and 80% had a peripheral catheter; 98% were treated with intravenous fluids, and 96% were treated with intravenous ranitidine. On the basis of results of a descriptive study and knowledge of the epidemiologic features of B. pickettii, a provisional causal hypothesis was formulated: intravenous ranitidine was the source of the outbreak. As a control measure, it was advised to stop treatment with this drug. On the basis of results of logistic regression and the microbiological isolation of B. pickettii in an ampule of the drug, we concluded that intravenous ranitidine was the cause of the outbreak.
Materials and Methods
Case descriptions. An outbreak of 46 cases of nosocomial
infection by B. pickettii occurred between June and November
1993 in three hospitals in Madrid: Hospital Universitario San
Carlos (HUSC; 1,500 beds), Hospital Severo Ochoa (HSO; 425
beds), and Hospital General de M6stoles (HGM; 401 beds).
Microbiological method. The clinical samples were pro
cessed by means of standard procedures. The drug samples
were inoculated in hyperconcentrated nutrient broth, brain
heart infusion broth, or commercial hemoculture bottles. Iso
lates of B. pickettii from clinical or environmental samples
were identified on the basis of colony morphology and results
of gram staining, the oxidase test, and the API 20 NE identifi
cation system (bioMerieux, Marcy l'Etoile, France).
Identification of the causes. We conducted a case-control
study of this epidemic. Every patient for whom microbiological
culture was positive for B. pickettii was defined as a case.
The controls were randomly selected from patients who were
admitted to the same department where a case was detected
on the same day ( 2 days) that the sample from which
B. pickettii had been isolated was obtained. Two controls were
selected for each case from a group of possible controls. The
information on all cases was gathered 48 hours before the
sample positive for B. pickettii was taken.
Statistical methods. The crude magnitude of the risk (odds
ratio) for the possible evaluated factors was estimated. Stra
tified analysis was performed to determine if there were any
biases or interactions. The logistic regression model was ad
justed by the forward strategy. The 95% confidence intervals
for the odds ratios were calculated.
Outbreak description. A total of 46 cases of B. pickettii
infection were reported. The infections first developed in
HUSC (20 cases). Twenty-one (46%) of the cases occurred in
HSO, and five cases (11 %) occurred in HGM. There was no
history of epidemiologic outbreaks due to B. pickettii in any
of the three hospitals. The epidemic curve (figure 1) suggested
that the cases were produced by a common source in HUSC
and HSO. The clinical characteristics of the cases are summa
rized in tables 1 and 2.
~'):~'S",'}4 :~~~~::t:~:~q~~~~~~,. "I~ 't/"fI, l 't/tO'I'JV fI,"
fI,(i,sJ~,;)(1' ...... '11 o'jV),"~ ,.}> 0,/'"v'" ..~."-_y~v"~-.4' .. "~.',~\-i!\:
Clinical/microbiological description. The 46 cases had 48
infections. The most frequent infection was bacteremia (fre
quency, 96% [44 cases]). Two (4%) of the cases had catheter
infections without associated bacteremia. Thirty-five percent
of the cases presented with an infection by another organism.
No death was attributed to the infection. Afterward, B. pickettii
was isolated from six parenteral nutrition samples that con
tained ranitidine (BBL, Alonga, Madrid). Ranitidine vials were
collected again for culture, and B. pickettii was isolated from
an intravenous ranitidine vial (lot R30; BBL, Alonga) at RUSe.
B. pickettii was isolated in 47 blood cultures, from three intra
venous catheter tips, and from an exudate of surgical drainage
of a pancreatic pseudocyst. Thirty-three isolates were analyzed
for the biotype.
Identification ofthe causes. The 92 controls were homoge
neous with the cases on all studied variables. The crude odds
ratios for the risk factors studied are shown in figure 2. When
Mean age (SO) in y
Previous average hospital stay (SO) in d
Total average hospital stay (SO) in d
Percentage of cases with indicated
Nonfatal basic disease
Department of Medicine
Recovery in ICU
NOTE. HGM = Hospital General de M6stoles; HSO = Hospital Severo
Ochoa; HUSC = Hospital Universitario San Carlos; ICU = intensive care
stratified analysis was applied to all variables that were signifi
cant in the univariate analysis, intravenous ranitidine remained
significant. We also observed that nasogastric tubes behaved
as a modifying variable with regard to the effect of intravenous
ranitidine (table 3).
There was epidemiologic evidence that B. pickettii was the
causal agent in our outbreak. Our findings were consistent with
those of all other reported outbreaks of B. pickettii infection
[2 -1 0]. Our outbreak involved a greater number of cases than
did any previously documented outbreak. B. pickettii variant
2, which was isolated in our outbreak, has been described as
the least virulent variant. In our study, this low virulence was
confirmed by the absence of death attributed to this microor
The greatest difficulty in the search for the cause of the
outbreak was the distribution of the drug in the hospitals. RUSe
used three different manufacturer's products for intravenous
ranitidine treatment, and the commercial brand of ranitidine
Coefficient OR 95% CI
Central venous catheter
Figure 2. Results of a univariate analysis
of the risk factors (top) and administered
treatments (bottom) of Burkholderia pickettii
infection. Crude odds ratios and 95% confi
dence intervals are shown.
that was administered was not recorded in the nurse's notebook.
HSO used only ranitidine from BBL, Alonga; this product was
suspected when B. pickettii was isolated from the parenteral
nutrition containing intravenous ranitidine. All the patients
from HSO with B. pickettii- induced bacteremia had been
treated with intravenous ranitidine within the previous 48
hours. The two patients with nasogastric tubes did not receive
intravenous ranitidine during the 48 hours before the specimen
for hemoculture was obtained. Therefore, the presence of a
nasogastric tube interacts with ranitidine administration. HSO
supplied intravenous ranitidine ampules to HGM, where a small
outbreak occurred; ranitidine was administered intravenously
in all five cases of the outbreak.
As a control measure, the boards of medical directors of
the three hospitals were informed when intravenous ranitidine
became suspected as the likely source of the outbreak (18
November 1993). The fact that no new cases were detected or
reported once the product stopped being dispensed (13 Decem
ber 1993) supports our hypothesis.
The intravenous ranitidine ampules may have been contami
nated during the sterilization process of its production. The
usual process consists of filtration through 0.45 or 0.22 micro
pore filters. Reports have shown that B. pickettii may be able
to pass through these filters . After the collection of the
drug vials, B. pickettii was isolated from two lots of the product
(H30 and H33).
Epidemiologic analysis and microbiological evidence (the
same biotype identified in all the strains studied and isolation
of B. pickettii from an ampule of the drug [lot H30]) confirmed
that intravenous ranitidine was the cause of nosocomial infec
tions by B. pickettii. This outbreak highlights the importance
of considering manufactured products as a possible cause of
nosocomial infections and offers a sobering reminder of the
need to maintain high standards for sterilization techniques and
quality control procedures.
The authors are grateful to A. Vicente, M.D., and M. Vinuesa,
M.D. (Preventive Medicine Service, Hospital General de Mos
toles); J. L. Gomez, M.D. (Microbiology Service, Hospital General
de Mostoles); R. Coello, M.D., M. Tafalla, M.D., P. Arroyo, R.N.,
D. Minguez, R.N., P. Uribe, R.N., and P. Sanchez, R.N. (Preven
tive Medicine Service, Hospital Universitario San Carlos); and P.
Fragoso, R.N., and M. J. Vega, R.N. (Preventive Medicine Service,
Hospital Severo Ochoa) for their cooperation and assistance during