The use of initial dosing of gentamicin in the management of pyelonephritis/urosepsis: A retrospective study
The use of initial dosing of gentamicin in the management of pyelonephritis/urosepsis: A retrospective study
Silvia Ryanto 0
Mandy Wong 0
Petra Czarniak 0
Richard Parsons 0
Katherine Travers 1
Matthew Skinner 1
Bruce SunderlandID 0
Baochuan Lin, Defense Threat Reduction
Agency, UNITED STATES
0 School of Pharmacy, Faculty of Health Sciences, Curtin University , Perth, Western Australia , Australia
1 Sir Charles Gairdner Hospital , Nedlands, Western Australia , Australia
The primary objective was to determine if initial empirical intravenous dosing of gentamicin
improved patient?s outcomes in pyelonephritis/urosepsis compared with alternative IV
Retrospective cross-sectional descriptive study.
Public TertiaryTeaching Hospital serving adults in an urban centre.
Competing interests: The authors have declared
that no competing interests exist.
in the study.
All adult patient records with a recorded diagnosis of any of pyelonephritis/urosepsis, urinary
tract infection, UTI, complicated urinary tract infection, bacteriuria, symptomatic bacteriuria
and asymptomatic bacteriuria from 2nd February 2012 to 10th May 2014 were reviewed.
Only patients treated with an empirical regimen of one or more IV antibiotics were included
The primary outcomes were: duration of IV antibiotic treatment, time to resolution of
symptoms and length of hospital stay (LOS). Secondary end points were: compliance with
Therapeutic Guidelines: Antibiotic (electronic version) (eTG) for severe pyelonephritis/urosepsis
and appropriateness of gentamicin use.
Univariate and multivariable associations between baseline and demographic variables and
the main outcomes were performed using Chi-square tests and a General Linear Model
using the SAS version 9.2 software.
Of 295 patients reviewed 152 were prescribed one or more IV antibiotics and included in the
study. Approximately half of the patients (n = 73, 48%) were prescribed IV
piperacillin/tazobactam (Tazocin), while gentamicin was prescribed for 66 patients (43.4%). Of the 152
patients evaluated, 49 (32%) were given gentamicin first. Multivariable regression analysis
showed that duration of IV treatment was shorter for those aged over 70 (40.2 hours vs 85.5
hours for those aged up to 70; p = 0.0074), and those who received gentamicin as first line
treatment (41.3 hours vs 89.8 hours for those not receiving any gentamicin; p = 0.0312).
After adjustment for age and gentamicin administration, there appeared to be no significant
difference in duration of IV treatment for any other independent variables. No significant
associations between the independent variables (gentamicin, age, gender, comorbidities,
and eTG compliance) and either time to resolution of symptoms (median: 68 hours) or
hospital LOS (median: 5 days) were observed.
Neither time to resolution of abnormal symptoms nor length of stay were significantly
influenced by an initial dose of gentamicin when the data were subjected to multivariable
analysis. The age of the patient and pattern of gentamicin treatment were the dominant factors
associated with duration of IV antibiotics. Piperacillin/tazobactam is not recommended in
treatment guidelines for pyelonephritis/urosepsis but was the most commonly prescribed IV
antibiotic. This requires review by the appropriate hospital clinicians.
Urinary tract infections (UTIs) are common in the community and hospital settings [
Approximately 0.7% of community presentations are for UTIs and they are the fourth most
common infection with a prevalence of 12.9%. Infection of the lower urinary tract is classified
as simple cystitis [
]; but with kidney involvement pyelonephritis [
]. Uncomplicated cystitis
more commonly occurs in young and sexually active women [
]. Pathogens from simple
cystitis may invade the systemic circulation, which can progress to pyelonephritis/urosepsis
]. The risk of this progression is increased by the presence of comorbidities including
metabolic disorders (such as diabetes), functional or structural abnormalities (such as infected
cysts, calculi, renal/bladder abscesses or renal stones) or unusual pathogens (such as yeasts)
]. The most common pathogen isolated is Escherichia coli (E. coli), accounting for 70?95% of
all UTI cases [
Diagnosis of a UTI can be difficult in the elderly as distinguishing between asymptomatic
bacteriuria and symptomatic infection especially when communication is poor may cause
delays and therefore increase severity of the disease.
2 / 11
Sepsis is an inflammatory response to a systemic infection [
], which can lead to
multiple organ failure. This is a serious condition, with a high mortality risk [
urosepsis accounts for approximately 5?7% of all sepsis cases, [
] and has a 20?42% mortality
]. As disease progression can occur quickly, immediate (within the first hour) IV
antibiotic therapy can improve patient outcomes [
Use of aminoglycosides has shown marked improvement in patient outcomes for infections
involving some common Gram-negative organisms such as E. coli, Klebsiella pneumoniae and
Pseudomonas aeruginosa [
]. In Australia, the Therapeutic Guidelines: Antibiotic
(electronic version) (eTG) , recommend gentamicin IV 4?6 mg/kg as a single dose with
amoxicillin/ampicillin IV 2 g 6 hourly as first line empirical therapy in urosepsis [
It was hypothesized that the initial administration of one or two IV doses of gentamicin as
part of empirical therapy for severe pyelonephritis/urosepsis would result in improved patient
outcomes. Therefore, the primary objective of this study was to determine whether patients
receiving gentamicin in addition to other IV antibiotics showed improved outcomes
compared with those receiving IV antibiotics in the absence of gentamicin. In addition the study
aimed to determine the level of appropriate use of gentamicin in the selected population, with
respect to standard treatment guidelines.
A retrospective, cross-sectional descriptive study was conducted involving patients admitted
for more than 24 hours to the Medical Assessment Unit (MAU) at a large 600 bed tertiary
teaching hospital in Perth Western Australia, between February 2, 2012, and May 10, 2014.
Staff at the hospital compiled a list of adult patients (male or female) whose diagnoses included
any of the following: pyelonephritis/urosepsis, urinary tract infection (UTI), complicated
urinary tract infection/UTI, bacteriuria, symptomatic bacteriuria and asymptomatic bacteriuria.
The initial diagnosis would have usually been made by hospital emergency doctors. The ICD
10 codes were searched electronically and medical records of these patients were made
available to researchers SR and MW. The study was approved by The Human Research Ethics
Committee at Curtin University (Protocol Approval PH-12-14) and the Sir Charles Gairdner
Hospital Ethics Committee with Quality Improvement (QI) Number 4559.
Adult male and female patients initially treated with an empirical regimen of one or more IV
antibiotics for the above diagnoses were included in the study. Patients who were given only
oral antibiotics or no antibiotics were excluded. Data were collected from patient hospital
records and included patient demographics (age, gender, weight, co-morbidities and whether the
patient was a nursing home resident), penicillin allergies, primary and secondary diagnosis,
symptoms on admission, and baseline signs (temperature, blood pressure, pulse).
Co-morbidities for pyelonephritis/urosepsis were defined as any risk factors that contributed to the
development of UTI, such as diabetes, indwelling catheter use, recurrent UTI, or renal obstructions.
The total number of IV antibiotics, total duration of treatment, time and dates of each
antibiotic regimen initiated and ceased, dose of each antibiotic, details of gentamicin dosing if given
(dose, times and date), discharge medication regimen (if given) were also obtained from the
patient medical records. Blood or urine microorganism identification and antibiotic sensitivity
data were obtained from the hospital database to identify the microorganism and antibiotic
suitability. Additional blood tests which included data on serum creatinine, gentamicin plasma
levels, white cell counts (WCC), and C-reactive protein (CRP) were also recorded.
3 / 11
The primary outcome was the time for resolution of pyelonephritis/urosepsis. Associated
measures for this outcome included the duration (hours) of IV antibiotic treatment, duration
(hours) to infection resolution and the hospital length of stay (LOS) (days). Time to resolution
was calculated from the medical record observational charts, based on the timing of the first
physiological parameter (temperature, pulse and blood pressure) indicating that the patient
had returned from ?abnormal? to ?normal?. Laboratory resolutions were defined as a WCC
normalised to 4?12 x 109/L and CRP decline of 50% from peak.
Secondary end points were the level of adherence with the eTG for the management of
severe pyelonephritis/urosepsis and the appropriateness of gentamicin use in this sample
population. Adherence to standard treatment guidelines was evaluated based on
recommendations in the eTG23, where first line treatment is stated (for the complete study period) as
gentamicin IV 4?6 mg/kg given as an immediate single dose with amoxicillin/ampicillin IV 2
g given every 6 hours [
]. For patients hypersensitive to penicillin, gentamicin alone can
be used for initial therapy. If gentamicin is contraindicated, the alternative is either IV
ceftriaxone 1 g or IV cefotaxime 1 g every eight hours . Where the patient?s weight was not
recorded and gentamicin was given, it was not possible to calculate whether the dose adhered
with the eTG, and this group was classified as non-adherent as the correct dose could not be
Descriptive statistics were used to summarise the main outcome variables according to patient
demographic and treatment profile (gentamicin treatment group, age, gender, co-morbidities,
eTG adherence [
]), the type of antibiotic prescribed and culture strains. Univariate
comparisons of the gentamicin treatment groups with respect to demographic variables were
conducted with Chi-square statistics. Univariate analyses to assess any relationships between each
of the main outcome measures: the duration of IV antibiotic treatment (hours), hospital LOS
(days) and time to resolution (hours), and the independent variables were performed.
Gentamicin dosing was classified into 3 groups as follows: 1) no gentamicin was given at all, 2)
gentamicin given first as empirical therapy, or 3) gentamicin given following other IV antibiotics.
Because the main outcome measures were not normally distributed (skewed), they were
subjected to a Box-Cox transformation to find the transformation that would bring the outcomes
closest to a Normal distribution. The logarithmic transformation was found to be optimum, so
the outcomes were log-transformed before analysis. A separate univariate general linear model
(GLM) for each of the main outcome measures was used to assess the statistical significance of
any differences between groups. Fisher?s Least Significant Difference (LSD) post-hoc analysis
was conducted to obtain the pairwise comparisons of the different gentamicin groups for each
outcome. Following univariate analysis, a multivariable analysis was performed for each
outcome, also using a GLM. All independent variables were initially included in each model, then
the least significant variable were dropped, one at a time, until all variables remaining in the
model were significantly associated with the outcome. The exception to this is that the variable
indicating the gentamicin group was forcibly included in the model, as this variable was of
primary interest. Pairwise interaction terms between main effects which remained in the model
were also assessed for significance. The resolution of symptoms on the basis of blood tests
were compared between gentamicin treatment groups using the Chi-square test. All statistical
analyses were conducted using the SAS version 9.2 software (SAS Institute Inc, Cary, NC,
USA, 2008), and a p-value 0.05 was taken to indicate a statistically significant association in
4 / 11
A total of 295 records of patients with a diagnosis of UTI were reviewed. Based on the
inclusion criteria, 152 with a diagnosis of severe pyelonephritis/urosepsis were included in the
study. The remaining 143 patients were excluded on the basis that their sole treatment for UTI
was an oral antibiotic (n = 127); the UTI was not treated at all (n = 6); the patient had other
concomitant infections (n = 9); or had been prescribed prophylactic antibiotics (n = 1).
Patients were predominantly female (n = 97, 64%), and the majority were over 70 years of age
(n = 80, 53%). In addition, 53 (35%) patients had existing co-morbidities.
The main microbiological cultures available from blood or urine samples are summarised in
Table 2. It was found that Escherichia coli (n = 65, 42.5%) was the most commonly isolated
pathogen, followed by Klebsiella pneumoniae and Pseudomonas aeruginosa (both n = 14, 9.2%
respectively). Other culture strains included: Citrobacter koseri, Proteus mirabilis, Klebsiella
oxytoca, Serratia marcescens, Morganella morganii and mixed growth.
Factors influencing patient outcomes
Review of the records showed that 31 (20%) adhered with guidelines. These included 11 (33%)
that were prescribed ceftriaxone IV. There were 9 cases where gentamicin was given first and
amoxicillin subsequently according to guidelines but the gentamicin dose could not be verified
as the patient?s weight was not recorded. Table 3 shows the distribution of the independent
variables (age, gender, comorbidities, eTG adherence across the gentamicin groups, with
Frequency of prescribing n = 152 n (%)
5 / 11
values to compare the profiles. The only significant association appeared to be with eTG
adherence, where this was significantly higher amongst those who received gentamicin first (33%)
than those who received it as a second line treatment (29%), or those who did not receive it at
Table 4 shows the descriptive statistics for the duration of IV antibiotic (hours), time to
resolution of symptoms (hours) and hospital LOS (days), within each of the groups defined by
the independent variables. The univariate GLM identified significant differences in the
duration of IV administration for the gentamicin group (p = 0.0248), for the age group
(p = 0.0056), and for the eTG adherence variable (p = 0.0204). Pairwise differences for the
gentamicin variable identified a significant difference in IV duration between patients given the
drug first vs not at all (p = 0.0073), with those receiving the drug first having a significantly
shorter duration of IV administration. Cases where the guidelines were followed also
experienced a shorter duration of IV administration than those where guidelines were not followed
(p = 0.0204).
While there appeared to be some differences in the time to resolution of symptoms, and
hospital LOS (Table 4), the only variable which reached statistical significance was the
gentamicin group and hospital LOS (p = 0.0496).
Multivariable analysis identified that only age group and gentamicin group were
significantly associated with the duration of IV administration (Table 5). The variable indicating
eTG adherence did not significantly contribute to this model, largely because eTG adherence
and gentamicin group were strongly correlated. The GLM analyses were performed on the
logarithm of the duration as there was considerable skewness in the raw measures. In addition to
the results of the regression analysis (exponential of the regression coefficients in the column
headed the ?ratio? and p-values), Table 5 also includes the adjusted means based on the original
scale (hours) for ease of interpretation. The values in the ?Ratio? column show the ratio of the
hours of each group relative to the reference group. While the adjusted means based on the
Gentamicin treatment group
raw data are not strictly accurate because of the skewness in the raw data, they give an
indication of the means on the original scale which may aid the understanding of the data. The mean
duration of IV treatment for those aged over 70 years was approximately 40 hours compared
with 85 hours for those up to the age of 70 (confidence intervals and p-values are shown in
Table 5). The duration of IV treatment for those who received gentamicin first was
significantly lower than those who did not receive gentamicin at all, but not significantly different
from those who received the drug second (based on the confidence intervals for the ratios).
Table 6 shows the numbers and percentages of patients with a recorded resolved CRP and
WCC upon discharge for each sub-category of gentamicin administration. The p-values
(obtained from a Fisher?s Exact test because numbers in some cells were small) suggest that
gentamicin administration was not associated with CRP and WCC resolution in this study.
There also appeared to be no association between the pattern of gentamicin use and renal
95% CI for B
Adjusted Mean (hours)
(95% CI )
This study has investigated empirical IV antibiotic treatment prescribed for 152 patients for
severe pyelonephritis/urosepsis with one or more IV antibiotics. The overall duration of IV
antibiotic treatment was significantly shorter for patients given gentamicin empirically as
initial treatment compared to patients not given gentamicin at all. eTG adherence was a factor in
reducing IV antibiotic duration but did not influence the other outcome parameters. Duration
of IV therapy is an important indicator as patients usually need to cease IV antibiotics before
discharge can be considered. The large spread of the data as can be seen from the wide ranges
possibly limited other significant findings.
This result suggested that an initial dose of IV gentamicin positively affected patient
outcomes in terms of IV antibiotic treatment length and hospital LOS. Gentamicin has long been
the antibiotic of choice in the treatment of severe cases of suspected gram-negative sepsis due
to its effectiveness against Pseudomonas aeruginosa [
]. Some patients were infected with
Pseudomonas aeruginosa. However E. coli was the main pathogen, a finding which has
previously been documented [
]. A recent study reported 54% of E.coli strains found in urine
were sensitive to gentamicin [
]. Adhering to eTG requirements also produced improved
outcomes for duration of IV antibiotics, but this was correlated with the pattern of gentamicin
administration which was the dominant factor. These data also included the initial
administration of ceftriaxone.
A limitation in this study was the imprecise recording of the diagnosis. For this study, the
initial prescribing of an IV antibiotic in the treatment regimen for ?UTI? was considered an
indication of severe pyelonephritis/urosepsis. Patients who were only given oral or no
antibiotics, were excluded as this suggested a diagnosis of acute cystitis or mild pyelonephritis as
opposed to severe pyelonephritis/urosepsis. The combination of the imprecise diagnosis and
some missing data for various parameters contributed to some limitations in the scope of data
analysis. In addition, the wide ranges of observed LOS, duration of treatment and time to
resolution, combined with the small to moderate sample size mean that the study may be
susceptible to Type II errors (true associations may have been missed).
Age was analysed against duration of IV antibiotic treatment, duration to disease resolution
and hospital LOS. Notably the older group (> 70 years) had a significantly shorter duration of
8 / 11
IV antibiotic treatment possibly owing to the awareness of increased risk of toxicity, such as
], in the elderly population.
Adherence with eTG [
] recommendations was low with 31 (20%) of IV antibiotics
prescribed in accordance with recommendations. This was partially caused by a lack of recording
of patient weight which then did not permit verification of the dose of gentamicin and these
were classified as non-adherent. Prescribing IV ceftriaxone would have also influenced the
eTG adherence outcome. In several cases IV ceftriaxone was initially prescribed followed by
IV gentamicin. This may have been once urinary excretion had been determined. A high
number of empirical prescriptions for piperacillin/tazobactam was found, or gentamicin followed
by piperacillin/tazobactam which are not treatment options under the guidelines. A study
conducted by Khan et al.[
], noted the significant widespread and inappropriate use of this
antibiotic in the hospital setting [
]. It is noted that piperacillin/tazobactam is not a restricted
antibiotic at the study hospital. Hence such an antibiotic can be prescribed without permission
even when not recommended in guidelines. However even when an antibiotic is effective
against the pathogen detected, it does not mean it is the most appropriate.
The lack of adherence to the eTG [
] recommendations could be due to a lack of
awareness of the current guidelines, uncertainty of differential diagnosis and complex
] (such as renal failure that could contraindicate first line empirical gentamicin
treatment as stated in the eTG [
]). However, it has been well documented that
aminoglycosides, such as gentamicin, have consistently demonstrated bactericidal activity as well as a post
antibiotic effect, hence they continue to manage both Gram-positive and Gram-negative
bacterial growth after they have been ceased [
15, 31, 32
]. Lee et al. [
] also reported that guideline
non-concordant empirical antibiotic treatment of acute pyelonephritis produced a lower early
clinical response rate and increased hospital LOS.
A secondary aim of this study was to determine the appropriateness of the use of
gentamicin in this population. Unfortunately, reliable conclusions could not be drawn, as crucial data,
such as patients? weight and renal function were inconsistently recorded. However, with the
available data, a chi-squared analysis was conducted on renal clearance and gentamicin dosing.
It was found that 25 out of 86 patients (29.1%) who were not given gentamicin at all had an
eGFR > 60mL/min, and this was not significantly different from the clearance rate for those
patients who were taking gentamicin (Table 6). Gentamicin also had no significant effect on
either CRP or WCC resolution times. These findings could have been partially influenced by
missing blood test results for a number of patients, in particular those admitted for two days or
less. As a result, these patients were unable to be included in that analysis, limiting any
conclusions that may be drawn. Although it could be possible that gentamicin was reserved for less
sick patients (and hence shorter duration of treatment and stay in itself) the renal function
data (influenced by age and illness) show no such correlation.
Further limitations of this study include the retrospective nature of the study (no control
over the treatment allocated to each patient), which may affect comparability between groups.
Some patient documentation was inconsistent and some important data were missing from
some records. These included the patient weight (which is an important determinant of
appropriateness of gentamicin dosing), renal function, and blood test results (in particular CRP and
WCC). The study was conducted at a single hospital which may limit the generalisability of the
findings. Often a general coded ?UTI? diagnosis was recorded rather than specifically cystitis
or pyelonephritis/urosepsis. This approach has been adopted since this study investigated
initial empirical prescribing rather than when a final diagnosis was made. Scrutiny of patient data
was necessary and 143 patients were rejected, to identify those initially prescribed IV
antibiotics. It is therefore possible that some confounding with respect to indication could have arisen,
but those patients were more likely to have been excluded rather than included.
9 / 11
There were univariate associations indicating that an initial dose of gentamicin significantly
improved patient outcomes in terms of duration of IV treatment, and this association persisted
after adjustment for age.
Failure to follow guidelines may be due to a lack of awareness of the current eTG
recommendations or caution in empirically prescribing gentamicin in the study population. As such,
it would be beneficial in future to evaluate prescribers? views regarding the current standard
treatment guidelines. The role of piperacillin/tazobactam as initial IV antibiotic management
in pyelonephritis/urosepsis requires review, especially in the light of these findings.
Conceptualization: Katherine Travers, Matthew Skinner.
Data curation: Silvia Ryanto, Petra Czarniak, Katherine Travers, Matthew Skinner.
Formal analysis: Silvia Ryanto, Mandy Wong, Richard Parsons.
Investigation: Mandy Wong, Bruce Sunderland.
Methodology: Silvia Ryanto, Mandy Wong, Petra Czarniak, Richard Parsons, Katherine
Travers, Matthew Skinner, Bruce Sunderland.
Project administration: Petra Czarniak.
Supervision: Petra Czarniak, Richard Parsons, Bruce Sunderland.
Validation: Silvia Ryanto, Petra Czarniak.
Writing ? original draft: Silvia Ryanto, Mandy Wong.
Writing ? review & editing: Petra Czarniak, Richard Parsons, Katherine Travers, Matthew
Skinner, Bruce Sunderland.
10 / 11
1. Tandogdu Z , Wagenlehner FM . Global epidemiology of urinary tract infections . Curr Opin Infect Dis . 2016 ; 29 ( 1 ): 73 - 9 . https://doi.org/10.1097/QCO.0000000000000228 PMID: 26694621
2. Johansen T , Cek M , Naber K. Prevalence of Hospital-Acquired Urinary tract infection in urology departments . Eur Urol . 2007 ; 51 ( 4 ): 1100 - 1111 . https://doi.org/10.1016/j.eururo. 2006 . 08 .012 PMID: 17049419
3. Hooton T. Uncomplicated urinary tract infection . N Engl J Med . 2012 ; 366 : 1208 .
4. Gupta K , Hooton T , Naber K. International clinical practice guidelines for the treatment of acute uncomplicated cystitis and pyelonephritis in women: a 2010 update by the Infectious Diseases Society of America and the European Society for Microbiology and Infectious Diseases . Clin Infect Dis . 2011 ; 52 ( 2 ): 103 - 120 .
5. Rahn D. Urinary tract infections: contemporary management . Urol Nurs . 2008 ; 28 : 333 - 341 . PMID: 18980099
6. Foxman B , Barlow R , D'Arcy H , Gillespie B , Sobel J . Urinary tract infections: self-reported incidence and associated costs . Ann Epidemiol . 2000 ; 10 : 209 - 615 .
7. Foxman B . Epidemiology of urinary tract infections: incidence, morbidity, and economic costs . Am J Med . 2002 ; 113 : 5 - 11S .
8. Wagenlehner F , Pilatz A , Naber K , Weidner W. Therapeutic challenges of urosepsis . Eur J Clin Invest . 2008 ; 38 ( 2 ): 45 - 49 .
9. Nicolle L . A practical approach to the management of complicated urinary tract infection . Drugs and Aging . 2001 ; 18 : 243 - 254 . https://doi.org/10.2165/ 00002512 -200118040-00002 PMID: 11341472
10. Czaja C , Scholes D , Hooton T . Population-based epidemiologic analysis of acute pyelonephritis . Clin Infect Dis . 2007 ; 45 ( 3 ): 273 - 280 . https://doi.org/10.1086/519268 PMID: 17599303
11. Echols R , Tosiello R , Haverstock D . Demographic, clinical, and treatment parameters influencing the outcome of acute cystitis . Clin Infect Dis . 1999 ; 29 ( 1 ): 113 - 119 . https://doi.org/10.1086/520138 PMID: 10433573
12. Balk R . Severe sepsis and septic shock. Definitions, epidemology, and clinical manifestations . Crit Care Clin . 2000 ; 16 : 179 - 192 . PMID: 10768078
13. Bone R , Balk R , Cerra F. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis . Chest . 1992 ; 101 : 1644 - 1655 . PMID: 1303622
14. Brun-Buisson C. The epidemiology of the systemic inflammatory response . Intensive Care Med . 2000 ; 26 ( 1 ): 64 - 74 .
15. Cunha B. Aminoglycosides : Current role in antimicrobial therapy . Pharmacotherapy . 1988 ; 8 : 334 - 350 . PMID: 3146747
16. Hotchkiss R , Karl I. The pathophysiology and treatment of sepsis . N Engl J Med . 2003 ; 348 : 138 - 150 . https://doi.org/10.1056/NEJMra021333 PMID: 12519925
17. Brunkhorst F. Epidemiology, economy and practice-results of the German study on prevalence by the competence network sepsis (Sep-Net) . Anasthesiol Intensivmed Notfallmed Schmerzther 2006 ; 41 : 43 - 44 . https://doi.org/10.1055/s-2005 -921227 PMID: 16440262
18. Martin G , Mannino D , Eaton S , Moss M. The epidemiology of sepsis in the United States from 1979 through 2000 . N Engl J . 2003 ; 348 : 1546 - 1554 .
19. Kreger B , Craven D , Carling P . Gram-negative bacteremia . III. Reassessment of eitology, epidemiology and ecology 612 patients . Am J Med . 1980 ; 68 : 1589 - 1596 .
20. Kreger B , Craven D , McCabe W . Gram-negative bacteremia. Re-evaluation of clinical features and IV treatment in 612 patients . Am J Med . 1980 ; 68 : 344 - 355 . PMID: 6987871
21. Wagenlehner F , Weidner W , Naber K. Optimal management of urosepsis from the urological perspective . Int J Antimicrob Agents . 2007 ; 30 ( 5 ): 390 - 397 . https://doi.org/10.1016/j.ijantimicag. 2007 . 06 .027 PMID: 17728107
22. Johnson J , Hardin T . Aminoglycosides, imipenem, and aztreonam. Clin Podiatr Med Surg . 1992 ; 9 : 443 - 464 . PMID: 1586911
23. Antibiotic . In: eTG complete. Melbourne: Therapeutic Guidelines Limited; [cited 23 /10/2014].
24. Doiron D , Fiebig DG , Suziedelyte A . Hips and hearts: The variation in incentive effects of insurance across hospital procedures . J Health Econ . 2014 ; 37 : 81 - 97 . https://doi.org/10.1016/j.jhealeco. 2014 . 06 .006 PMID: 24981504
25. Hooton T. Clinical practice. Uncomplicated urinary tract infection . N Engl J Med . 2012 ; 366 ( 11 ): 1028 - 1037 . https://doi.org/10.1056/NEJMcp1104429 PMID: 22417256
26. Holloway W , Taylor W. Gentamicin and kanamycin in the treatment of Gram-negative sepsis: A comparative study . J Infec Diseases . 1971 ; 124 : S180 - S184 .
27. Holloway W. Gentamicin : clinical and pharmacological studies . Int J Clin Pharm . 1970 ; 3 : 122 - 127 .
28. Akinbowale O , Peng H , Barton M. Antimicrobial resistance in bacteria isolated from aquaculture sources in Australia Journal Of Applied Microbiology . 2006 ; 100 ( 5 ): 1103 - 1113 . https://doi.org/10. 1111/j.1365- 2672 . 2006 . 02812 . x PMID : 16630011
29. Fraisse T , Aygon C , Paccalin M , Vitrat V , DeWazieres B , Baudoux V . Aminoglycosides use in patients over 75 years old . Age and Aging . 2014 ; 43 : 676 - 681 .
30. Khan F , Elhiday A , Khudair I , Yousef H , Hussein A . Evaluation og the use of piperacillin/tazobactam (Tazocin) at Hamad General Hospital, Qatar: are there unjustified prescriptions? Infection and Drug Resistance . 2012 ; 5 : 17 - 21 . https://doi.org/10.2147/IDR.S27965 PMID: 22294859
31. Barriere S. Aminoglycosides : A reassessment of their therapeutic role . Clin Pharm . 1988 ; 7 : 385 - 390 . PMID: 3289824
32. Zhanel G , Craig W. Pharmacokinetic contributions to post-antibiotic effects . Clin Pharmacokinetics . 1994 ; 27 : 377 - 392 .
33. Lee SS , Kim Y , Chung DR . Impact of discordant empirical therapy on outcome of community-acquired bacteremic acute pyelonephritis . J of Infection 2011 ; 62 : 159 - 164 .