Community antibiotic prescription and pneumonia hospital case fatality rate in elderly
EDITOR- A recent meta-analysis by D’Amico et al. reports that
antibiotic prophylaxis can reduce respiratory tract infections and overall
mortality in critically ill patients.* Despite this published evidence of
efficacy, population based evidence of effectiveness of antibiotic use in
reducing subsequent mortality for pneumonia is scarce. Pneumonia is one
of the five major causes of death in the elderly of industrialised
countries. As an example, in Friuli Venezia Giulia, North-East Italy,
pneumonia in the elderly accounted for 2.3% of total death in 1989 - 1997
and for 3.6% of total discharges from public hospital during 1997. To
investigate the possible effect of community prescribed antibiotics on the
risk of hospital death among elderly patients with pneumonia, we conducted
a computerised record-linkage study utilising the databases of
dispensed pharmaceutical prescriptions and hospital admissions of Friuli-
Venezia Giulia from November 1, 1995 to April 30, 1996. The cases were all
the residents in the region, aged > 65 discharged from public
hospitals with a principal diagnosis of pneumonia (ICD-9: 480 - 487). For
these subjects a complete history of antibiotic prescribing
within 30 days prior to hospital admission was ascertained.
Among the 315,108 local elderly, 136,583 scripts were cashed for
antibiotics, during 1996. Quinolones accounted for 26%, macrolides for 25%
penicillins for 22 %, cephalosporins for 18% and sulfamides-cotrimoxazole
for 6%. 1296 cases (median age 82 years) were identified. The fatality
rate was 22.1%. Co-morbidity was primarily represented by cardiac disease
(69.5%), chronic
obstructive pulmonary disease(COPD)(22.9%), diabetes mellitus (12.4%),
neoplasms (6.7 %) and liver diseases (6.4%). 54.6% of the cases had cashed
at least one antibiotic prescription: quinolones (16.5%), penicillins
(12.4%), macrolides (11.6%). Use of antibiotics during the 30 days
previous to hospitalisation in elderly was associated with approximately a
50% decreased mortality due to pneumonia. The best fitted multivariate
logistic model (goodness of fit P>chi2=0.94, ROC area 0.76) showed
that previous use of quinolones, macrolides or penicillins was
independently associated with a significant reduction in the risk of death
(Table).
Although we can not exclude confounding by indication as an explanation of
the association found in the analysis, we controlled our results for
length of hospital stay and source of referral, which are proxies of
severity of illness. Community use of antibiotics in patients
subsequently hospitalised with pneumonia could be either truly protective
against hospital deaths (0.25 overall attributable prevention) or a marker
of community acquired pneumonia as distinguished from nosocomial
pneumonia.
*D’Amico R, Pifferi S, Leonetti C, Torri V,Tinazzi A, Liberati A:
Effectiveness of antibiotic prophylaxis in critically ill adult patients:
systematic review of randomised controlled trials. BMJ 1998;316:1275-85
OR for deaths in hospitalised cases of pneumonia in elderly,
according to a backward stepwise logistic regression model, robust
standard errors & 95%CI adjusted for clustering on the regional
hospitals included in the analysis.
OR for deaths in hospitalised cases of pneumonia in elderly, according to
a backward stepwise logistic regression model, robust standard errors
& 95%CI adjusted for clustering on the regional hospitals included
in the analysis.
Rapid Response:
Community antibiotic prescription and pneumonia hospital case fatality rate in elderly
EDITOR- A recent meta-analysis by D’Amico et al. reports that
antibiotic prophylaxis can reduce respiratory tract infections and overall
mortality in critically ill patients.* Despite this published evidence of
efficacy, population based evidence of effectiveness of antibiotic use in
reducing subsequent mortality for pneumonia is scarce. Pneumonia is one
of the five major causes of death in the elderly of industrialised
countries. As an example, in Friuli Venezia Giulia, North-East Italy,
pneumonia in the elderly accounted for 2.3% of total death in 1989 - 1997
and for 3.6% of total discharges from public hospital during 1997. To
investigate the possible effect of community prescribed antibiotics on the
risk of hospital death among elderly patients with pneumonia, we conducted
a computerised record-linkage study utilising the databases of
dispensed pharmaceutical prescriptions and hospital admissions of Friuli-
Venezia Giulia from November 1, 1995 to April 30, 1996. The cases were all
the residents in the region, aged > 65 discharged from public
hospitals with a principal diagnosis of pneumonia (ICD-9: 480 - 487). For
these subjects a complete history of antibiotic prescribing
within 30 days prior to hospital admission was ascertained.
Among the 315,108 local elderly, 136,583 scripts were cashed for
antibiotics, during 1996. Quinolones accounted for 26%, macrolides for 25%
penicillins for 22 %, cephalosporins for 18% and sulfamides-cotrimoxazole
for 6%. 1296 cases (median age 82 years) were identified. The fatality
rate was 22.1%. Co-morbidity was primarily represented by cardiac disease
(69.5%), chronic
obstructive pulmonary disease(COPD)(22.9%), diabetes mellitus (12.4%),
neoplasms (6.7 %) and liver diseases (6.4%). 54.6% of the cases had cashed
at least one antibiotic prescription: quinolones (16.5%), penicillins
(12.4%), macrolides (11.6%). Use of antibiotics during the 30 days
previous to hospitalisation in elderly was associated with approximately a
50% decreased mortality due to pneumonia. The best fitted multivariate
logistic model (goodness of fit P>chi2=0.94, ROC area 0.76) showed
that previous use of quinolones, macrolides or penicillins was
independently associated with a significant reduction in the risk of death
(Table).
Although we can not exclude confounding by indication as an explanation of
the association found in the analysis, we controlled our results for
length of hospital stay and source of referral, which are proxies of
severity of illness. Community use of antibiotics in patients
subsequently hospitalised with pneumonia could be either truly protective
against hospital deaths (0.25 overall attributable prevention) or a marker
of community acquired pneumonia as distinguished from nosocomial
pneumonia.
*D’Amico R, Pifferi S, Leonetti C, Torri V,Tinazzi A, Liberati A:
Effectiveness of antibiotic prophylaxis in critically ill adult patients:
systematic review of randomised controlled trials. BMJ 1998;316:1275-85
OR for deaths in hospitalised cases of pneumonia in elderly,
according to a backward stepwise logistic regression model, robust
standard errors & 95%CI adjusted for clustering on the regional
hospitals included in the analysis.
OR for deaths in hospitalised cases of pneumonia in elderly, according to
a backward stepwise logistic regression model, robust standard errors
& 95%CI adjusted for clustering on the regional hospitals included
in the analysis.
Deaths % Survivals % OR* (95%CI) Other respiratory disease 44 15.3 109 10.8 1.6 1.1-2.4 COPD 22 7.7 227 22.5 0.4 0.3-0.7 Cerebrovascular diseases 13 4.5 14 1.4 3.7 1.9-7.2 Severe associated symptoms 37 12.9 25 2.5 7.6 4.7-11.1 Penicillins 28 9.8 132 13.1 0.6 0.4-0.9 Macrolides 23 8.0 127 12.6 0.5 0.3-0.8 Quinolones 28 9.8 186 18.4 0.4 0.3-0.6*Odds ratio adjusted for age, sex, co-morbidity, length of hospital
stay and source of referral.
Competing interests: Deaths % Survivals % OR* (95%CI)Other respiratory disease 44 15.3 109 10.8 1.6 1.1-2.4COPD 22 7.7 227 22.5 0.4 0.3-0.7Cerebrovascular diseases 13 4.5 14 1.4 3.7 1.9-7.2Severe associated symptoms 37 12.9 25 2.5 7.6 4.7-11.1Penicillins 28 9.8 132 13.1 0.6 0.4-0.9Macrolides 23 8.0 127 12.6 0.5 0.3-0.8Quinolones 28 9.8 186 18.4 0.4 0.3-0.6