braz j infect dis 2 0 1 9;2 3(4):237–245
The Brazilian Journal of
INFECTIOUS DISEASES
www.elsevier.com/locate/bjid
Original article
Clinical and economic impact of generic versus
brand name meropenem use in an intensive care
unit in Colombia
Karen Ordóñez a , Max M. Feinstein b,c,e , Sergio Reyes b,e ,
Cristhian Hernández-Gómez b,e , Christian Pallares b,d,e , María V. Villegas
b,d,e,∗
a
E.S.E. Hospital Universitario San Jorge, Pereira, Colombia
Centro Internacional de Entrenamiento e Investigaciones Médicas (CIDEIM), Cali, Colombia
c Case Western Reserve University, Cleveland, United States
d Centro Médico Imbanaco, Cali, Colombia
e Grupo de Investigación en Resistencia Antimicrobiana y Epidemiología Hospitalaria, Universidad El Bosque, Bogotá, Colombia.
b
a r t i c l e
i n f o
a b s t r a c t
Article history:
Background: Recent studies suggest that sustained use of generic antibiotics may be asso-
Received 17 March 2019
ciated with clinical failure and emergence of antibacterial resistance. The present study
Accepted 16 June 2019
was designed to determine the clinical outcome between the use of generic meropenem
Available online 22 July 2019
(GM) and brand-name meropenem (BNM). Additionally, this study evaluated the economic
impact of GM and BNM to determine if the former represents a cost-effective alternative to
Keywords:
the latter.
Generic drugs
Methods: Patients treated between January 2011 and May 2014 received GM while patients
Meropenem
treated between June 2014 and March 2017 received BNM. Mortality was compared between
Mortality
groups. Total infection cost was defined by the cost of antimicrobial consumption, length of
Costs
stay, and laboratory and imaging exams until infection resolution.
Critical care
Findings: A total of 168 patients were included; survival rate for the 68 patients treated
Gram negative bacteria
with GM was 38% compared to 59% in the patients treated with BNM. Multivariate analysis showed that the variables most strongly-associated with mortality were cardiovascular
disease (OR 18.18, 95% CI 1.25–262.3, p = 0.033) and treatment with generic meropenem (OR
18.45, 95% CI 1.45–232.32, p = 0.024). On the other hand, total infection cost did not show a
significant difference between groups (BNM $10,771 vs. GM $11,343; p = 0.91).
Interpretation: The present study suggests that patients treated with GM have a risk of death
18 times higher compared to those treated with BNM. Furthermore, economic analysis
shows that GM is not more cost effective than BNM.
∗
Corresponding author at: Research Group on Antimicrobial Resistance and Hospital Epidemiology-RAEH, Universidad El Bosque,
Bogotá, Colombia. Mailing address: Cra 9 # 131 A- 02, Lab de Investigacion 2 Piso, Bogotá, Colombia.
E-mail addresses: karenmelissao@gmail.com (K. Ordóñez), maxfeinstein@gmail.com (M.M. Feinstein), sergiorsor1990@gmail.com (S.
Reyes), crihergo@gmail.com (C. Hernández-Gómez), icako@hotmail.com (C. Pallares), mariavirginina.villegas@gmail.com (M.V. Villegas).
https://doi.org/10.1016/j.bjid.2019.06.010
1413-8670/© 2019 Sociedade Brasileira de Infectologia. Published by Elsevier España, S.L.U. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
238
b r a z j i n f e c t d i s . 2 0 1 9;2 3(4):237–245
Summary: More studies measuring clinical outcomes are needed to confirm the clinical
equivalence of brand-name versus generic antibiotics, not only for meropenem but also for
other molecules.
© 2019 Sociedade Brasileira de Infectologia. Published by Elsevier España, S.L.U. This is
an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).
Introduction
The generic medication contains the same components and
specifications as the brand-name medication. In Colombia,
generic medications are available for purchase 10 years after
the commercial patent of the original brand-name medication expires. Generic medications must comply with a variety
of conditions to be approved for commercial use. Broadly
speaking, generic medications must contain the same quantitative and qualitative composition of active ingredients
compared to their brand-name counterpart. Furthermore,
the pharmaceutical presentation must be similar to their
respective brand-name counterpart with regards to formulations (e.g., parenteral, enteral, etc.), safety profile, and
pharmacokinetics.1 The prescription of generic medications
is considered one of the most effective approaches for reducing cost and increasing global access to first-line medication
for the treatment of diverse infections.2,3
Many countries have progressively opened their markets
to generic medications. In 2010, generic antibiotics (GA) represented more than two-thirds of the global consumption of
medications.3 In the United States, generic medications made
up to 63% of total prescriptions in 2007,4 while in England
that number was 83% in 2008.5 The generic medication market has likewise expanded in other developed countries such
as Japan.2
Recently, the use of generic medications has caused controversy, particularly in the case of GA. It has traditionally been
considered that GAs, principally in parental form, have the
same efficacy as brand-name antibiotics (BNA).2,3,6–9 However,
several studies have questioned the clinical effectiveness of
GAs, reporting issues like clinical failure or emergence of bacterial resistance associated with the sustained use of Gas.3,5,7,8
Antibiotic resistance has also been related to a significant use of GAs following their release into the drug market.
This collateral damage has been demonstrated in Germany
and Denmark where use of generic quinolones generated a
disproportionate increase in its use, followed by an increase
in resistance levels from 7.7% to 25% and 0.8% to 4%,
respectively.10,11 This phenomenon has also been reported
in Colombia where the use of -lactams has surpassed the
90th percentile of the defined daily dose, considerably more
than has been reported in other countries.12–14 On a national
level, some antibiotics are consumed more than others, as
is the case with meropenem given the ready availability of
its generic version. Meropenem provides adequate coverage
against Pseudomonas aeruginosa and is considered first-line
treatment for a number of complicated infections including
intra-abdominal infections, bacterial meningitis, blood stream
infections, nosocomial pneumonia, septicemia, and febrile
neutropenia.15
On the other hand, the characteristics of GAs can be
affected by a variety of elements, such as active ingredient,
bioequivalence, impurities, additives, and excipients. Many of
these elements depend on the manufacturing process, which
can vary among GAs and BNAs. In a country where GAs
represent a large portion of medications consumed due to
limited resources, the therapeutic efficacy of GAs needs to
be evaluated. This will contribute to greater confidence in the
prescription of GAs, as well as provide a better understanding
of possible deleterious effects related with their use.
Keeping in mind the current uncertainties regarding possible differences between GAs and BNAs, this study was
designed to determine the clinical impact associated with
the use of generic versus brand-name meropenem for the
treatment of infections caused by Gram-negative bacteria susceptible to carbapenems. This study evaluated patients who
were hospitalized in the intensive care unit (ICU) of a tertiary
care hospital in Colombia.
Materials and methods
Study design
This cohort study compared patients who received generic
and brand-name meropenem. The group that received
generic meropenem included patients hospitalized in the ICU
between January 2011 and May 2014 in a tertiary care hospital in Pereira, Colombia. The introduction of the generic was
part of the hospital administration strategy to reduce costs. In
contrast, patients hospitalized from June 2014 to March, 2017
in the same ICU were treated with brand-name meropenem
based on a new policy of buying brand antibiotics again; these
patients constitute the brand-name cohort.
During both time periods of the study, the infection control surveillance did not change nor the prevention strategies
implemented by the Infectious Disease Committee. Also, the
dosing, infusion and time of treatment for meropenem was
the same based on the institutional antimicrobial guidelines
in both groups.
Inclusion criteria for the study were age over 18 years,
patients hospitalized in the ICU, an infection caused by
meropenem-susceptible Gram-negative bacteria, and treatment with meropenem. Patients were excluded if any of the
following criteria were met: death within the first 72 h of initiation of meropenem therapy, a concomitant invasive fungal
infection, concomitant prescription of four or more doses of
antibiotics with the same antibacterial activity as meropenem,
and incomplete patient records.
For sample size calculation, mortality in Colombian ICUs
associated to intrahospital infections was assumed to be
b r a z j i n f e c t d i s . 2 0 1 9;2 3(4):237–245
25.6%.16–20 This proportion was considered for unexposed
patients, that is, those patients who received brand-name
meropenem. The difference between the proportion among
exposed and unexposed patients was considered; for the
exposed patients who received generic medication, an
increase in mortality of 68.5%21 was assumed. A statistical
power of 80% and a confidence level of 95% were used, with
an exposed/unexposed ratio of 1 and a loss of 10%. The calculated sample size was 140 patients in each arm, totaling 280
patients. Exposed and unexposed were matched for age (in
10-year ranges) and for type of infection. Sampling was performed by including non-probabilistic patients consecutively.
Sample size was calculated in the program EPIDAT version 3.1.
Because we were unable to complete the expected sample size,
our power decreased 26.6%.
The study was conducted under the appropriate local
and international ethics guidelines and approved by the
institutional review board of the International Center for
Training and Medical Research (CIDEIM, per its abbreviation in Spanish) and the participating institution. CIDEIM
deemed that informed consent was not necessary for the
study given that its design entailed minimal risk for the subjects.
Data collection
Patient data relevant to the study was retrieved from the
hospital’s electronic medical record system and recorded on
written data collection forms. Data collected included patient
demographic information, hospital admission and discharge
dates, ICU admission and discharge dates, clinical and laboratory data needed to calculate Sequential Organ Failure
Assessment (SOFA) score, comorbidities, infection type, susceptibility profile of the infecting organism, antibiotic doses
received during the infection episode, clinical and laboratory
tests ordered between the infection episode and discharge,
and survival status upon discharge.
Statistical analysis
In the descriptive statistical analysis, proportions were
determined for qualitative variables, and mean or median
calculated as determined by the distribution of quantitative
variables. The odds ratio for mortality associated with the
use of generic and brand-meropenem was computed using
two-by-two contingency tables. Chi square or Fisher exact
tests were used for comparisons of qualitative variables. For
quantitative variables, parametric or non-parametric tests
were used depending on whether distribution was normal or
not. Survival analysis for 7- and 28-day survival both cohorts
were compared using Log-rank test, adjusted for severity and
adequate therapy. Both cohorts were matched for age and
type of pathology. In the multivariate analysis, odds ratio
was calculated using logistic regression with adjustment for
the matching variables, and a conditional logistic regression was performed for death as an outcome. All possible
models included paired variables, starting with those with a
p-value <0.20 in the univariate analysis. In the multivariate
analysis, p-value <0·05 was considered statistically significant.
239
Economic analysis
The cost of care for patients with infections that were selected
for the study corresponded to the sum of the antibiotic consumption and length of stay until infection resolution. A
partial economic assessment was carried out using microcosting techniques as a way of determining the magnitude of
the resources spent. Allocation of costs for each resource was
based on a reference cost and adjusted by the inflation rate
based on the Colombian manual document for costs.22 Incremental cost-effectiveness ratio (ICER), which is a measure
of cost-effectiveness for a given intervention, was calculated
using decision tree model. Survival was defined as the clinical outcome for effectiveness, while economic outcome was
defined as the total infection cost per patient.
Results
Clinical impact
A total of 1289 patients received meropenem during the
study period, 318 patients received generic meropenem and
971 received brand-name meropenem. Of the 318 patients
treated with generic meropenem, 68 (21.3%) met inclusion
criteria for the study. Of the 971 patients treated with brandname meropenem, 100 (10.3%) met these criteria. Thus, a
total of 168 patients were included in the study. Demographic
characteristics such as SOFA score, duration of antibiotic treatment, time of infusion and dosing of meropenem, as well as
the microorganisms causing the infection were comparable
among groups (Table 1). Regarding infection type, the most
common was ventilator-associated pneumonia in the generic
group (23% vs. 8%, p = 0.005) and abdominal infection in the
brand-name group (13% vs. 34%, p = 0.004). On the other hand,
the brand-name group had a higher prevalence of comorbidities such as type 2 diabetes (15% vs. 28%, p = 0.043) and
cardiovascular disease (23% vs. 38%, p = 0.049). Also, history of
hospitalization in the three months prior to hospital admission was more common in the brand-name group (53% vs. 72%,
p = 0.011).
We categorized the SOFA in lower and higher than 10, since
the SOFA greater than 10 is associated with a mortality of more
than 40%. On day 7 and 28 there were no differences in mortality by severity. In contrast, there was higher mortality at day 7
and 28, in patients using generic meropenem (Fig. 1).
Mortality was greater in the generic group (62% vs. 41%,
p = 0.008); univariate analysis was performed to assess risk
factors associated with mortality. Statistically significant variables were age greater than 73 (RR 4.98, 95% CI 1.31–18.96,
p = 0.018), history of cardiovascular disease (RR 2.53, 95% CI
1.23–5.26, p = 0.02), presence of central venous catheter (RR
2.92, 95% CI 1.07–8.74, p = 0.02), ventilator-associated pneumonia (RR 6.43, 95% CI 1.99–26.9, p = 0.000), and treatment with
generic meropenem (RR 2.32, 95% CI 1.18–4.59, p = 0.008), as
shown in Table 2.
As both generic and brand-name meropenem groups were
comparable by age and comorbidities in previous analyses, the
multivariate analysis was adjusted for SOFA, GLASGOW and
240
b r a z j i n f e c t d i s . 2 0 1 9;2 3(4):237–245
Table 1 – Univariate analysis.
Variable
Sex
Male
Female
Age (years)
18–28
29–39
40–50
51–61
62–72
>72
Age
Previous hospitalization
Comorbidities
Diabetes
Pulmonary disease
Immunosuppression
Neurologic disease
Renal disease
Liver disease
Cardiovascular disease
Cancer
Sequential Organ Failure Assessment (SOFA) score
0–6
7–9
10–12
13–14
SOFA
Glasgow Coma Scale (GCS) score
Mild
Moderate
Severe
Glasgow
Invasive devices
Orotracheal tube
Central venous catheter
Foley catheter
Infection type
Bloodstream infection
Ventilator-associated pneumonia
Urinary tract infection
Skin and soft tissue infection
Meningitis
Intraabdominal infection
Dosing interval of meropenem
6 hours
8 hours
12 hours
24 hours
Prolonged meropenem infusion
Time between culture and initiation of meropenem treatment
Immediate
24–72 hours
>72 hours
already receiving
Time between culture and initiation of meropenem
Days of treatment with meropenem
<7 days
7 days
Ceftriaxone
Ciprofloxacin
Multidrug resistant bacteria
Appropriate therapy
Generic meropenem group
n = 68 (41%)
Brand name meropenem group
n = 100 (59%)
p- value
42 (62%)
26 (38%)
54 (54%)
46 (46%)
7 (10%)
7 (10%)
7 (10%
12 (18%)
14 (21%)
21 (31%)
58.6 (SD ± 19.7)
36 (53%)
6 (6%)
16 (16%)
11 (11%)
24 (24%)
19 (19%)
24 (24%)
56.47(SD ± 17.45)
72 (72%)
10 (15%)
8 (12%)
3 (4%)
7 (10%)
6 (9%)
1 (1%)
16 (23%)
5 (7%)
28 (28%)
16 (16%)
7 (7%)
2 (2%)
4 (4%)
1 (1%)
38 (38%)
3 (3%)
10 (40%)
7 (28%)
6 (24%)
2 (8%)
7.72 (SD = 2.96)
17 (52%)
11 (33%)
5 (15%)
0 (0%)
6.69 (SD = 2.59)
43 (65%)
7 (11%)
16 (24%)
12.37 (SD ± 3.64)
54 (69%)
6 (8%)
18 (23%)
12.11 (SD ± 4.25)
62 (91%)
60 (88%)
59 (87%)
85 (85%)
83 (83%)
86 (86%)
0.235
0.349
0.887
29 (43%)
16 (23%)
9 (13%)
3 (4%)
2 (3%)
8 (13%)
54 (54%)
8 (8%)
10 (10%)
4 (4%)
1 (1%)
33 (34%)
0.149
0.005
0.516
1
0.565
0.004
0.290
1 (2%)
60 (88%)
7 (10%)
0 (0%)
4 (10%)
1 (1%)
94 (94%)
3 (3%)
2 (2%)
23 (24%)
26 (38.2%)
10 (14.7%)
12 (17.6%)
1 (1.4%)
18 (SD ± 39.01)
19 (73%)
1 (4%)
1 (4%)
0
43.7 (SD ± 53.13)
1/49 (2%)
22/50 (44%)
16/51 (31%)
18/53 (34%)
15 (22%)
66 (97%)
0/14 (0%)
13/21 (62%)
9/23 (39%)
6/22 (27%)
20 (20%)
99 (99%)
0.318
0.440
0.011
0.043
0.441
0.742
0.032
0.319
1
0.049
0.271
0.180
0.660
0.096
0.765
0.778
0.168
0.514
0.572
0.747
0.351
241
b r a z j i n f e c t d i s . 2 0 1 9;2 3(4):237–245
– Table 1 (Continued)
Variable
Infectious organism
Escherichia coli
Klebsiella pneumoniae
Enterobacter spp.
Proteus spp.
Pseudomonas aeruginosa
Serratia marcescens
Acinetobacter baumannii
Morganella morganii
Mortality
Generic meropenem group
n = 68 (41%)
Brand name meropenem group
n = 100 (59%)
18 (27%)
16 (24%)
7 (11%)
1 (1%)
8 (12%)
3 (4%)
10 (15%)
3 (4%)
42 (62%)
41 (43%)
16 (17%)
5 (5%)
2 (2%)
18 (19%)
6 (6%)
5 (5%)
3 (3%)
41 (41%)
0.008
7-Day survival analysis by exposure (p =0,0054)
0.00
0.00
0.25
0.25
0.50
0.50
0.75
0.75
1.00
1.00
7-Day Life Analysis by SOFA SEVERITY (p = 0.64)
p- value
2
4
6
2
0
8
0
SOFA SCORE < 10
4
6
8
analysis time
analysis time
Brand-name meropenem
SOFA SCORE >/= 10
Generic Meropenem
28-Day Survival analysis by exposure (p=0,0045)
10
analysis time
0
SOFA SCORE < 10
20
30
0
0.00
0.00
0.25
0.25
0.50
0.50
0.75
0.75
1.00
1.00
28-Day Survival analysis by SOFA severity (p = 0,83)
10
20
30
40
50
analysis time
Brand-name meropenem
SOFA SCORE >/= 10
Generic Meropenem
Fig. 1 – Life charts.
time of administration of meropenem. Exposure and cardiovascular disease were statistically significant.
Patients who received generics had 18-fold higher risk of
dying compared with those who received brand-name (OR:
18.45 95%CI: 1.46–232) and patients with cardiovascular disease also had 18-fold higher risk of dying compared with those
who did not have this comorbidity (OR: 18.1 95%CI: 1.26–262).
In spite of the wide confidence intervals, due the small sample
size, they did not include 1, with a good statistical significance
(Table 3).
Economic impact
The total antimicrobial cost was lower in the brand-name
group as compared to the generic group ($303 vs. $588). The
total cost of ICU stay was also lower in the brand-name group
($8,896 vs. $7,705). However, the total cost of infection was
not significantly different between groups (brand-name cost
$10,771 vs. generic cost $11,343).
The ICER is a measure of cost-effectiveness for a given
intervention, which in this case is the choice of antibiotic. The
242
b r a z j i n f e c t d i s . 2 0 1 9;2 3(4):237–245
Survival Rate
Brand-name
Meropenem
(100)
59%
Total Infection
Cost
Cost difference
$10,771 USD
Patients with Gram
negative infections
$572 USD
Generic
Meropenem
(68)
38%
Survival
difference
21%
Incremental
costeffectiveness
ratio (ICER)
$2,724 USD
$11,343 USD
Fig. 2 – Cost-effectiveness decision tree model.
ICER was defined by the cost of obtaining one additional effectiveness unit (patient survival) was $2,724 USD per ICU stay
when changing from brand name molecule to generic. The
cost-effectiveness decision tree model is shown in Fig. 2.
Discussion
In this study, we compared critically ill patients treated
with generic versus brand-name meropenem. Results demonstrated greater mortality in the generic group and comparable
total costs for both groups. History of cardiovascular disease
was also found to be an independent risk factor for mortality.
In contrast, the brand-name had a higher prevalence of comorbidities and history of hospitalization in the three months
prior to hospital admission.
Meropenem is one of the most commonly-used antibiotics
in critically ill patients due to its broad-spectrum bactericidal
activity and its pharmacokinetic characteristics. The present
study suggests clinical superiority of brand-name meropenem
with similar costs in care. There are few clinical studies that
compared clinical outcomes with treatment of generic versus
brand-name antibiotics. Most studies have been conducted
in vitro and animal models.23,24 Also, systematic reviews on
the efficacy and quality of GAs have not been able to conclude their inferiority in relation to BNAs, maybe in part due
to the heterogenous nature and different end points of the
studies included, which precluded a definitive conclusion.25
Furthermore, it is difficult to have the possibility of evaluating
brand versus generic meropenem in the same hospital, under
the same conditions, like we were able to do in our study. For
example in Thailand, a study comparing generic versus brandname meropenem in hospitalized patients, of which 60% had
documented microbiological data and some were infected
with meropenem-resistant bacteria, did not find inferiority of
either molecule.24 However, in our study the inclusion criteria
was meropenem susceptible Gram-negative bacteria, eliminating the possibility that resistance could interfere with
outcome ascertainment in the two groups. Another Colombian prospective cohort study that included 1015 patients
with nosocomial infections showed in multivariate analysis
a risk of death almost two-fold higher in patients treated
with GAs compared to patients treated with BNAs (HR = 1.91;
95%CI = 1.43–2.55).21
Several studies led by Vesga et al. have compared GAs and
BNAs using in vivo models. One of these studies using three
different brands of generic vancomycin showed lower in vivo
effectiveness compared to brand-name vancomycin despite
similarities of chemical equivalence and potency. This difference was statistically significant and independent of route
of administration or doses.7 Furthermore, Vesga et al. have
shown difference in efficacy with other antibiotics, including meropenem, in which a pharmacologic equivalence did
not translate into therapeutic equivalence in animal infection
models.9 In this study, the authors proposed that differences
in meropenem efficacy might be secondary to trisodium compounds that increase susceptibility in the face of hydrolysis
within the organism. Finally, Vesga et al. also evaluated the
impact of GA use on bacterial susceptibility profiles.8 An animal model was used to evaluate the resistance profile of
Staphylococcus aureus in the face of successive cycles of GA and
BNA. It was observed that generic vancomycin progressively
selected subpopulations of resistant bacteria, confirming the
effect of suboptimal bactericidal action of GA on microorganism susceptibility. In contrast to the Vesga et al. findings, some
authors like Louie et al. and Hadwiger et al. have not been
able to replicate these results.26,27 Irrespective of the contradictory findings with animal models, our study results show a
statistically significant difference in the clinical outcomes.
Another key point is the therapeutic indications for GA.
In the U.S., the Food and Drug Administration therapeutic indications for generic and brand-name meropenem are
identical. However, the Colombian National Institute of Food
and Drug Surveillance Agency (INVIMA in Spanish) has recommended only brand-name meropenem for the treatment
of febrile neutropenia rather generic meropenem. Likewise,
INVIMA recommends only brand-name vancomycin for the
treatment of pneumonia, sepsis, or meningitis caused by
penicillin-resistant Streptococcus pneumoniae; difference in the
therapeutic indications between the GA versus the BNAs
raises concerns whether GAs can be prescribed in all kinds
of patients, especially when critically ill.
This study is different from other studies regarding the
clinical outcomes between generic versus brand-name antibiotics. First, the selected study population was limited to
critically ill patients. This population was selected because
their conditions demand that antibiotics used have good
b r a z j i n f e c t d i s . 2 0 1 9;2 3(4):237–245
Table 2 – Bivariate analysis.
243
Table 3 – Multivariate analysis.
Variable
Relative risk (95% CI /p-value)
Variable
Sex
Age
18–28
29–39
40–50
51–61
62–72
73+
Comorbidities
Diabetes
Pulmonary disease
Immunosuppression
Neurologic disease
Renal disease
Cardiovascular disease
Solid Organ Tumor
Previous hospitalization
Sequential Organ Assessment Failure
(SOFA) score
0–6
7–9
10–12
13–14
Glasgow Coma Scale (GCS) score
Mild
Moderate
Severe
Invasive devices
Endotracheal tube
Central venous catheter
Foley catheter
Infection type
Bloodstream infection
Vent.-assoc. pneumonia
Urinary tract
Skin and soft tissue
Meningitis
Intraabdominal
Time b/w culture and initiation of
meropenem treatment
Immediate
<24 hours
24–72 hours
72+ hours
Obtained prior to ICU
Multidrug resistant bacteria
Use of generic meropenem
Infectious organism
P. aeruginosa
Enterobacteria
0.72 (0.37–1.38/p = 0.285)
Sequential organ failure assessment
score
(Group 1)
(Group 2)
(Group 3)
Glasgow score
(Group 1)
(Group 2)
Treatment with generic meropenem
Comorbidities
Diabetes
Cardiovascular disease
Infection type
Bloodstream
Vent.-assoc. pneumonia
Urinary tract
Intraabdominal
Time b/w culture and initiation of
meropenem treatment
(Group 1)
(Group 2)
(Group 3)
(Group 4)
Solid organ tumor
(Reference group)
0.47 (0.10–2.34/p = 0.359)
1.80 (0.40–8.07/p = 0.443)
2.51 (0.65–9.67/p = 0.180)
2.39 (0.61–9.33/p = 0.210)
4.98 (1.31–18.96/p = 0.018)
1.79 (0.88–4.03/p = 0.119)
1.25 (0.48–3.30/p = 0.614)
1.03 (0.23–4.64/p = 0.969)
0.81 (0.16–3.92/p = 0.760)
1.02 (0.22–4.64/p = 0.969)
2.53 (1.23–5.26/p = 0.006)
3.23 (0.55–33.48/p = 0.138)
1.19 (0.60–2.35/p = 0.597)
(Reference group)
1.96 (0.58–6.61/p = 0.276)
0.71 (0.17–3.03/p = 0.648)
1.25 (0.07–22.13/p = 0.879)
(Reference group)
0.95 (0.30–3.03/p = 0.931)
1.40 (0.64–3.08/p = 0.397)
2.75 (0.94–9.09/p = 0.041)
2.92 (1.07–8.74/p = 0.020)
1.62 (0.61–4.52/p = 0.289)
0.56 (0.29–1.08/p = 0.064)
6.43 (1.99–26.90/p = 0.000)
1.88 (0.64–5.96/p = 0.203)
1.38 (0.23–9.73/p = 0.676)
2.1 (0.11–125.33/p = 0.539)
0.45 (0.20–1.00/p = 0.034)
(Reference group)
0.78 (0.31–1.96/p = 0.597)
1.13 (0.43–3.02/p = 0.800)
0.96 (0.33–2.83/p = 0.941)
0.88 (0.40–1.96/p = 0.760)
1.28 (0.57–2.91/p = 0.516)
2.32 (1.18–4.59/p = 0.008)
1.57 (0.61–4.20/p = 0.303)
0.64 (0.24–1.63/p = 0.303)
effectiveness and safety due to severity of illness. Critical
patients usually have alterations in the volume of distribution,
tissue perfusion and renal functions among others, which
may lead to significant challenges in achieving recommended
pharmacokinetic and pharmacodynamic parameters for optimal treatment against bacterial infections.28 Second, only
patients with documented Gram-negative infections susceptible to meropenem were included to maximize the likelihood of
successful treatment with the selected antibiotics. To reduce
confounding factors in the mortality analysis, patients with
invasive fungal infections were excluded. Third, the economic
analysis allowed the comparison of cost of hospitalization
Odds Ratio (95% CI /p-value)
1.07 (0.15–7.50/p = 0.943)
0.73 (0.08–6.62/p = 0.776)
0.64 (0.02–25.33/p = 0.812)
2.76 (0.20–37.98/p = 0.449)
31.33 (0.60–1639.36/p = 0.088)
18.45 (1.45–232.32/p = 0.024)
2.93 (0.40–21.38/p = 0.288)
18.18 (1.25–262.63/p = 0.033)
3.40 (0.27–41.61/p = 0.338)
5.92 (0.16–208.88/p = 0.328)
2.49 (0.07–82.87/p = 0.609)
4.54 (0.25–79.87/p = 0.301)
0.64 (0.06–6.81/p = 0.717)
0.27 (0.02–2.97/p = 0.288)
0.05 (0.00–4.09/p = 0.189)
1.15 (0.09–14.71/p = 0.910)
3.11 (0.08–107.83/p = 0.530)
as well as the calculation of incremental cost of survival for
each treatment used. Fourth, the groups of patients treated
with generic and brand-name meropenem were comparable
in terms of age and severity of infections. Furthermore, the
group of patients treated with the brand-name molecule had
a higher prevalence of cardiovascular disease, which was a
significant risk factor for mortality.
Pharmacoeconomic analyses are important for an evaluation of costs and benefits of health interventions. In the
present study, a cost-effectiveness model was created to
compare costs when using each molecule. The net healthcareassociated costs were included in the assessment of each
molecule, as well as the secondary costs of complications
resulting from therapeutic failure and adverse reactions associated with the antibiotic. On the other hand, the cost-of-stay
in the ICU was higher in the generic group due to the
development of more complications and unfavorable clinical
progression. At the end, there were no significant differences
in the costs of treatment for each episode of infection between
cohorts, which is the most cogent argument for using generics.
When the incremental cost effectiveness ratio (ICER) was
calculated for the brand-name and generic molecules in our
study, the ICER analysis showed that treatment with brandname meropenem confers greater patient survival at a lower
cost and is therefore is a better treatment option than its
generic counterpart.
A limitation of this study is that only 67% of the anticipated sample size of 252 was achieved. This occurred primarily
due to the inclusion and clinical follow-up criteria, which was
strict and therefore led to the exclusion of many patients. Also,
secondary to the smaller sample size achieved, some variables
like SOFA or other non-cardiovascular comorbidities where
not found to be associated with mortality in the multivariate analysis. While a smaller sample may alter the precision
244
b r a z j i n f e c t d i s . 2 0 1 9;2 3(4):237–245
of the study, our results demonstrate a strong statistical association that is unlikely to be by chance. Even though the GA
had a higher incidence of ventilator associated pneumoniae
which may contribute in part to the higher mortality, the multivariate analysis did not show any difference associated with
this type of infection. Also, as we explained before, we did not
find any other factors that could explain the higher mortality
with GA like changes in the antibiotic prescription or infection
control practices.
In conclusion, the use of brand-name meropenem for
the treatment of Gram-negative infections susceptible to carbapenems in the ICU is a more cost-effective option than
generic meropenem. Institutions should follow clinical and
microbiologic outcomes of patients treated with antibiotics,
which in turn may allow for early detection of therapeutic failure. Ultimately, this approach can provide data that will help
drug producers improve the development of antibiotics.
editing. Address: Cra 9 # 131 A- 02, Lab de Investigacion 2 Piso,
Postal Code 110121, Bogotá, Colombia.
Christian Pallares designed the study, interpreted the data,
revised the manuscript, and approves of the final version.
Conceptualization, formal analysis, methodology, software,
validation, writing original draft, project administration, writing review and editing. Address: Av. La Maria #19-225, Postal
Code 760031, Cali, Valle del Cauca, Colombia.
María V. Villegas designed the study, revised the
manuscript, and approves of the final version. Conceptualization, funding acquisition, project administration, resources,
supervision, validation, visualization, writing-review and
editing. Address: Cra 9 # 131 A- 02, Lab de Investigacion 2 Piso,
Postal Code 110121, Bogotá, Colombia.
All authors agree to be accountable for all aspects of the
work and ensure the accuracy and integrity of the study.
references
Financial support
This study was funded by the International Center for Medical Training and Research (CIDEIM in Spanish). This study did
not receive any funding from drug manufacturers or any other
sources.
Conflicts of interest
The authors declare no conflicts of interest.
Thanks
We are grateful for the assistance we received from Obed David
Suárez Anaya, Carmen Elisa Llanos Uribe, Luz Stella Salazar,
and Camila Marin Peralta.
Author information
Karen Ordóñez designed the study, collected and interpreted
data, revised the manuscript, and approves of the final version.
Conceptualization, Data curation, Investigation, Methodology,
Supervision, Validation, Visualization, Writing original draft,
writing review and editing. Address: Carrera 18#12-75 Torre 1
Consultorio 1205, Postal Code 660003, Pereira, Colombia.
Max M. Feinstein collected and interpreted data, drafted
the manuscript, and approves of the final version. Address:
2058 E. 115th St., Cleveland, Ohio, Postal Code 44106, USA. Data
curation, formal analysis, writing original draft, writing review
and editing.
Sergio Reyes collected and interpreted data, revised the
manuscript, and approves of the final version. Data curation,
formal analysis, investigation, supervision, validation, writing original draft. Address: Av. La Maria #19-225, Postal Code
760031, Cali, Valle del Cauca, Colombia.
Cristhian Hernández-Gómez designed the study, interpreted the data, revised the manuscript, and approves of the
final version. Conceptualization, Formal analysis, investigation, methodology, software, supervision, validation, writing
original draft, project administration writing review and
[1]. U.S. Food And Drug Administration Code of Federal
Regulations Title 21.
[2]. Fujimura S, Watanabe A. Generic antibiotics in Japan. J Infect
Chemother. 2012;18(4):421–7.
[3]. Gauzit R, Lakdhari M. Generic antibiotic drugs: is
effectiveness guaranteed? Med Mal Infect. 2012;42(4):141–8.
[4]. Frank RG. The ongoing regulation of generic drugs. N Engl J
Med. 2007;357(20):1993–6.
[5]. Duerden MG, Hughes DA. Generic and therapeutic
substitutions in the UK: are they a good thing? Br J Clin
Pharmacol. 2010;70(3):335–41.
[6]. Mastoraki E, Michalopoulos A, Kriaras I, Mouchtouri E,
Falagas M, Karatza D, et al. Incidence of postoperative
infections in patients undergoing coronary artery bypass
grafting surgery receiving antimicrobial prophylaxis with
original and generic cefuroxime. J Infect. 2008;56(1):35–9.
[7]. Vesga O, Agudelo M, Salazar BE, Rodriguez CA, Zuluaga AF.
Generic vancomycin products fail in vivo despite being
pharmaceutical equivalents of the innovator. Antimicrob
Agents Chemother. 2010;54(8):3271–9.
[8]. Rodriguez CA, Agudelo M, Zuluaga AF, Vesga O. Generic
vancomycin enriches resistant subpopulations of
Staphylococcus aureus after exposure in a neutropenic mouse
thigh infection model. Antimicrob Agents Chemother.
2012;56(1):243–7.
[9]. Agudelo M, Rodriguez CA, Pelaez CA, Vesga O. Even
apparently insignificant chemical deviations among
bioequivalent generic antibiotics can lead to therapeutic
nonequivalence: the case of meropenem. Antimicrob Agents
Chemother. 2014;58(2):1005–18.
[10]. Kaier K, Frank U, Meyer E. Economic incentives for the
(over-)prescription of broad-spectrum antimicrobials in
German ambulatory care. J Antimicrob Chemother.
2011;66(7):1656–8.
[11]. Jensen US, Muller A, Brandt CT, Frimodt-Møller N,
Hammerum AM, Monnet DL. Effect of generics on price and
consumption of ciprofloxacin in primary healthcare: the
relationship to increasing resistance. J Antimicrob
Chemother. 2010;65(6):1286–91.
[12]. WHO Collaborating Center for Drug Statistics Methodology:
Definition and general considerations.
[13]. National Center for Infectious Diseases, Centers for Disease
Control and Prevention, Division of Healthcare Quality
Promotion, Public Health Service, US Department of Health
and Human Services. National Nosocomial Infections
b r a z j i n f e c t d i s . 2 0 1 9;2 3(4):237–245
Surveillance, Public Health Service, System Report data
summary from January 1992 through June 2004, issued
October 2004. Am J Infect Control. 2004;32:470–85.
[14]. Gutierrez GB. Relación entre el consumo de antibióticos y la
resistencia bacteriana en Bogotá. Univ Nac Colomb. 2009.
[15]. Baldwin CM, Lyseng-Williamson KA, Keam SJ. Meropenem: a
review of its use in the treatment of serious bacterial
infections. Drugs. 2008;68:803–38.
[16]. Pérez A, Dennis RJ, Rondón MA, Metcalfe MA, Rowan KM. A
Colombian survey found intensive care mortality ratios were
better in private vs. public hospitals. J Clin Epidemiol.
2006;59:94–101.
[17]. Dennis RJ, Pérez A, Rowan K, Londoño D, Metcalfe A, Gómez
C. Factors associated with hospital mortality in patients
admitted to the intensive care unit in Colombia. Arch
Bronconeumol. 2002;38:117–22.
[18]. Dennis R, Casas A, Urina M, Brainsky A, Rodríguez MN.
Predicción de mortalidad en cuidado intensivo Médicos,
Apache II y MPM. Acta Med Colomb. 1996;21:
17–26.
[19]. Díaz JG, Gómez Morales G, Reines B, Salomón A. Estudio
descriptivo sobre la correlación de veintiseís variables
demográficas y fisiológicas con la mortalidad de una
población de pacientes en una unidad de cuidado intensivo
en Colombia. Univ Med. 1994;35:7–15.
[20]. Durán Pérez J, Rodríguez García LC, Alcalá-Cerra G.
Mortalidad e infecciones nosocomiales en dos unidades de
cuidados intensivos de la ciudad de Barranquilla (Colombia).
Rev Científica Salud Uninorte. 2012;24.
245
[21]. Pallares CJ, Martínez E. Mortality risk factors associated with
healthcare infections in a tertiary level university hospital in
Colombia. Biomedica. 2014;34:148–55.
[22]. Colombia Digital Government—SOAT Manual; 2017.
[23]. Agudelo M, Rodriguez CA, Zuluaga AF, Vesga O. Relevance of
various animal models of human infections to establish
therapeutic equivalence of a generic product of
piperacillin/tazobactam. Int J Antimicrob Agents.
2015;45(2):161–7.
[24]. Angkasekwinai N, Werarak P, Chaiyasoot K, Thamlikitkul V.
Monitoring of effectiveness and safety of generic formulation
of meropenem for treatment of infections at Siriraj Hospital. J
Med Assoc Thai. 2011;94:S217–24.
[25]. Tattevin P, Crémieux AC, Rabaud C, Gauzit R. Efficacy and
quality of antibacterial generic products approved for human
use: a systematic review. Clin Infect Dis. 2014;58:458–69.
[26]. Louie A, Ii TB, Patel V, Huntley C, Liu W, Fikes S, et al.
Pharmacodynamic evaluation of the activities of six
parenteral vancomycin products available in the United
States. Antimicrob Agents Chemother. 2015;59:622–32.
[27]. Hadwiger ME, Sommers CD, Mans DJ, Patel V, Boyne MT.
Quality assessment of U.S. marketplace vancomycin for
injection products using high-resolution liquid
chromatography-mass spectrometry and potency assays.
Antimicrob Agents Chemother. 2012;56:2824–30.
[28]. Felton TW, Hope WW, Roberts JA. How severe is antibiotic
pharmacokinetic variability in critically ill patients and what
can be done about it? Diagn Microbiol Infect Dis [Internet].
2014;79:441–7.