Keywords cost-effectiveness - carbetocin - oxytocin - prevention postpartum hemorrhage - pregnancy
- high-risk
Palavras chave análise custo-eficiência - carbetocina - oxitocina - prevenção - hemorragia pós-parto
- gravidez de alto risco
Introduction
Postpartum hemorrhage (PPH) is defined as a blood loss greater than 500 mL within
24 hours postpartum and is considered severe if the loss reaches 1,000 mL in the same
time interval.[1 ] Postpartum hemorrhage affects ∼ 2% of women in labor and it is responsible a quarter
of maternal deaths. Uterine atony is the most common cause of this condition, but
there may be others, such as genital trauma, uterine rupture, placental retention
or maternal clotting disorders. The methods of active behavior during delivery, which
include administration of prophylactic uterotonics after the birth of the newborn,
early clamping and cutting of the umbilical cord, controlled traction of the cord,
and uterine massage, have been associated with a considerable reduction in the frequency
of PPH.[1 ]
Colombia has made progress in reducing maternal mortality by 25%; however, there are
still ∼ 400 deaths per year associated with pregnancy and delivery.[2 ] During 2012, when surveillance for the event of extreme maternal morbidity was initiated,
hemorrhagic complications ranked second, with 23.1%. In 2013, 2014, 2015 and 2016,
the proportions of cases of extreme maternal morbidity attributed to hemorrhagic complications
were 20%, 20.2%, 18% and 14.5%, respectively.[3 ] Likewise, there has been a decline in the proportion of maternal deaths attributed
to bleeding complications, reaching 13% in 2016.[3 ] These values show the impact of active behavior applied in the local context in
reducing maternal morbidity and mortality attributed to PPH.
Uterotonic agents for prevention of PPH include oxytocin, methylergonovine, misoprostol
and, more recently, carbetocin. Oxytocin is the most frequently used agent but, due
to its short half-life, it must be administered by intravenous infusion for sustained
uterotonic activity. Carbetocin has an efficacy and safety profile very similar to
oxytocin and, due to its pharmacokinetic characteristics, has a longer uterotonic
activity.[4 ]
[5 ]
[6 ] According to the national guidelines, oxytocin is the uterotonic agent of choice
in prevention; however, international guidelines like those from the Royal College
of Obstetricians and Gynaecologists (RCOG) and the Society of Obstetricians and Gynaecologists
of Canada (SOGC) already include carbetocin as an alternative, mainly in elective
caesarean section.[4 ]
[5 ]
[6 ]
The objective of this evaluation was to estimate the incremental cost-effectiveness
ratio of carbetocin compared with oxytocin for prevention of PPH by uterine atony
in women with at least one risk factor for hemorrhage, from the perspective of the
Colombian health system. Women with low risk for hemorrhage were not considered, since
the indication for them is less clear. The study was based on the guidelines established
in the manual for the elaboration of economic assessments in health proposed by the
national Institute of Technology Assessment in Health (IETS).[7 ]
Methods
A cost-effectiveness analysis was developed from the perspective of the third-party
payer to compare the use of carbetocin in the prevention of PPH with that of oxytocin
in women with the presence of at least one risk factor for hemorrhage due to uterine
atony. Among the risk factors described for PPH due to uterine atony are: multiple
gestation, polyhydramnios, macrosomia, large multiparous, severe hydrocephalus, prolonged
labor and chorioamnionitis.[8 ] A model for vaginal delivery and another one for cesarean delivery are presented
separately because the characteristics of these populations may change, as well as
risk factors and response to interventions. The main outcome was prevention of PPH,
defined as a blood loss exceeding 500 mL.[9 ] Also, as part of the sequential analysis, the consequences of the hemorrhagic event
with each of the alternatives were considered, particularly the treatment of PPH,
use of additional uterotonics and the requirement of surgical intervention. Considering
that there are several outcomes that measure the effectiveness of treatment (prevention
of PPH, use of additional uterotonics, adverse events), quality-adjusted life years
(QALYs) were used as a common outcome.[7 ] The time horizon considered in the economic evaluation was 1 year, so no discount
rate was required.
Decision Model
The models were constructed from the review of previous studies,[10 ]
[11 ]
[12 ]
[13 ]
[14 ] and were discussed and validated by a group of thematic experts. Two decision trees
were developed with TreeAge Pro Suite 2009 (TreeAge Software Inc., Williamstown, MA,
USA), one for cesarean section ([Fig. 1 ]) and one for vaginal delivery ([Fig. 2 ]). The model begins with women with the presence of at least one risk factor for
PPH due to uterine atony who can receive prophylaxis with any of the technologies
being evaluated: carbetocin or oxytocin. The first branch defines the occurrence or
not of PPH (blood loss > 500 mL), which corresponds to an analysis of the response
to preventive interventions. Carbetocin is available as an injectable solution of
100 mcg/mL and is a drug for hospital use that should be administered in a single
dose of 100 mcg in both vaginal and cesarean deliveries. Oxytocin is available in
an injectable solution of 5 IU/mL or 10 IU/mL, and the recommended dose in cases of
vaginal delivery is 5 to 10 IU, intramuscularly, but in case of cesarean section,
an IV bolus of 5 IU is recommended followed by an infusion of 30 IU.
Fig. 1 Decision model in the analysis of PPH in cesarean section.
Fig. 2 Decision model in the analysis of PPH in vaginal delivery.
Subsequently, sequential events that take place in the event of PPH occur until survival
or death at hospital discharge and whether or not the recommended treatments are given.
If PPH is present, the patient is given a first line of treatment, and if no response
is obtained, additional uterotonics are administered; if no response to the pharmacological
treatment is obtained, a first surgical approach including uterine tamponade and/or
hemostatic suture is performed. If the patient does not respond to this treatment,
a second-line surgical procedure would be performed, which usually corresponds to
hysterectomy.
Costs
Due to the perspective considered, only direct medical costs required to prevent and
treat hemorrhagic events with each of the alternatives were taken into account. All
costs were expressed in Colombian pesos in 2016. The resource identification was based
on the information from clinical practice guidelines and that provided by a group
of clinical experts.
We considered laboratories, procedures and hospital stay required in the process of
care of a patient from the administration of carbetocin or oxytocin. If no PPH was
present, the costs of the drugs administered in prevention and the hospital stay after
vaginal delivery and cesarean section were considered. If the event was present, the
costs of first and second line pharmacological treatments, and first and second line
surgical approaches were considered. To obtain the cost of medications, the following
sources were consulted: 2016 Report of the Drug Price Information System (SISMED)[7 ] and official documents of price regulation for medicines issued by the Ministry
of Health.[15 ] The average price reported for a carbetocin ampoule of 100 mcg was Col$ 95,800 and
that of the oxytocin ampoule of 10 IU was Col$ 1,130. The calculation of costs for
procedures associated in particular with the treatment follow-up was performed with
the Social Security Institute (ISS, in the Spanish acronym) Fee Manual (Agreement
256 of 2001), with a 30% increase for the case base scenario.[16 ] The prices of medical supplies, particularly surgical material, were obtained from
the “Colombia Efficient Acquisition” (Secop I, in the Spanish acronym) portal.[17 ]
Only the most frequently reported adverse events for carbetocin and oxytocin, which
are nausea and vomiting, were considered in the model. The identification of resources
required for the management of adverse reactions was performed with the assistance
of a clinical expert. The costs of adverse events were included within the treatment
costs, weighted by the frequency of presentation. The total costs in vaginal and cesarean
deliveries for the compared alternatives are presented in [Tables 1 ] and [2 ], respectively.
Table 1
Total costs in the vaginal delivery model (base case)
Group of patients
Carbetocin
Oxytocin
No PPH
$ 183,245
$ 89,442
PPH responding to the 1st line of pharmacological treatment
$ 1,426.303
$ 1,332.499
PPH that respond to additional uterotonics
$ 109,843
$ 16,040
PPH that respond to the first surgical approach
$ 472,038
$ 1,332.499
PPH responding to hysterectomy
$ 1,310.406
$ 1,216.603
Death
$ 231,494
$ 137,691
Abbreviation: PPH, postpartum hemorrhage.
Table 2
Total costs in the cesarean model (base case)
Group of patients
Carbetocin
Oxytocin
No PPH
$ 271,727
$ 177,134
PPH responding to the 1st line of pharmacological treatment
$ 1,427.365
$ 1,332.772
PPH that respond to additional uterotonics
$ 110,906
$ 16,613
PPH that respond to the first surgical approach
$ 964,178
$ 869,585
PPH responding to hysterectomy
$ 1,311.469
$ 1,216.876
Death
$ 231,468
$ 136,798
Abbreviation: PPH, postpartum hemorrhage.
Effects
The model parameters were obtained from randomized clinical trials comparing carbetocin
with oxytocin, from local databases and expert consultation. The studies considered
were identified from a panoramic review developed to evaluate the effectiveness and
safety of carbetocin compared with other oxytocics in the prevention of PPH.
To obtain information on mortality in patients with PPH, several sources were consulted:
the Individual Health Service Provision Registry (RIPS, in the Spanish acronym),[18 ] vital statistics of National Administrative Department of Statistics (DANE, in the
Spanish acronym),[19 ] the National Public Health Surveillance System (SIVIGILA, in the Spanish acronym)[3 ] and the published literature. Consultation in RIPS and vital statistics was performed
using the following diagnostic codes: O700, O701, O702, O703, O709, O712, O713, O714,
O719, O720, O721, O730, O731, O900, O901, O902, O723, O622 years O621. For the baseline
scenario, the mortality data for hemorrhagic complications were taken from SIVIGILA,
and a sensitivity analysis was performed considering a minimum of 0% and a maximum
of 3.1% as reported in the study by Sheldon et al.[20 ]
The quality-of-life (utility) weights were obtained through a search in the cost-effectiveness
analysis (CEA) Registry of Tufts University. Baseline utilities are those related
to vaginal and cesarean deliveries. From these, there is a decrease in the quality
of life in the population that presents PPH, either because of the requirement of
transfusions or because of the need for surgical interventions. The parameters considered
in the models are presented in [Table 3 ].
Table 3
Model parameters
Parameter
Base
Min
Max
Distribution
Source
Vaginal delivery
RR PPH carbetocin
0.950
0.430
2.090
Boucher et al. (2004)[21 ]
Probability of PPH in oxytocin group
0.164
0.074
0.361
Beta
Boucher et al. (2004)[21 ]
Probability of PPH in carbetocin group
0.156
0.071
0.343
Beta
Calculated
Probability of death by hemorrhagic event
0.003
0.000
0.031
Beta
INS[3 ]
Sheldon et al. (2014)[20 ]
RR use of additional uterotonics carbetocin
0.930
0.440
1.940
Boucher et al. (2004)[21 ]
Probability use of additional uterotonics oxytocin
0.156
0.074
0.325
Beta
Boucher et al. (2004)[21 ]
Probability use of additional uterotonics carbetocin
0.145
0.069
0.302
Beta
Calculated
Caesarean section
RR PPH carbetocin
0.520
0.290
0.930
Borruto et al. (2009)[22 ]
Probability of PPH in oxytocin group
0.442
0.247
0.791
Beta
Borruto et al. (2009)[22 ]
Probability of PPH in carbetocin group
0.230
0.128
0.411
Beta
Calculated
Probability of death by hemorrhagic event
0.008
0.000
0.031
Beta
INS[3 ]
Sheldon et al. (2014)[20 ]
RR use of additional uterotonics carbetocin
0.700
0.510
0.980
Borruto et al. (2009)[22 ]
Attilakos et al. (2010)[23 ]
Probability use of additional uterotonics oxytocin
0.374
0.272
0.524
Beta
Borruto et al. (2009)[22 ]
Attilakos et al. (2010)[23 ]
Probability use of additional uterotonics carbetocin
0.262
0.191
0.367
Beta
Calculated
Vaginal and cesarean delivery
Requirement for first-line surgical intervention (tamponade and hemostatic suture)
0.100
0.000
0.120
Beta
Experts
Requirement for second line surgical intervention (hysterectomy)
0.010
0.000
0.050
Beta
Experts
Surgical approach
Proportion of patients undergoing uterine tamponade
0.60
Experts
Proportion of patients undergoing hemostatic suture
0.40
Experts
Frequency of complications of uterine tamponade
Infection
0.02
Experts
Frequency of complications of hemostatic suture B-Lynch
Ureteral injury
0.002
Experts
Bladder injury
0.001
Experts
Intestinal injury
0.001
Experts
Frequency of hysterectomy complications
Infection
0.02
Experts
Ureteral injury
0.01
Experts
Bladder injury
0.01
Experts
Intestinal injury
0.005
Experts
Utilities
Vaginal delivery utility (Non-PPH)
0.9997
0.9996
0.9998
Triangular
[24 ]
Caesarean section utility (No PPH)
0.9996
0.9995
0.9997
Triangular
[24 ]
PPH (First Line Treatment) Utility - vaginal delivery
0.9918
0.9913
0.9922
Triangular
[25 ]
PPH (First Line Treatment) Utility - Caesarean section
0.9917
0.9912
0.9921
Triangular
[25 ]
Disutility by endometritis
0.0144
0.0132
0.0155
Triangular
[25 ]
Disutility due to operative injury
0.0303
0.0280
0.0326
Triangular
[25 ]
Hysterectomy disutility
0.0334
0.0306
0.0363
Triangular
[25 ]
Abbreviation: PPH, postpartum hemorrhage; RR, risk ratio; RR PPH, risk ratio of postpartum
hemorrhage.
Cost-effectiveness
The incremental cost was calculated as the difference between the costs, and the incremental
effectiveness as the difference between QALYs obtained when comparing the alternatives.
Incremental cost and incremental effectiveness were used to calculate the incremental
cost-effectiveness ratio (ICER): The ICER was compared with a cost-effectiveness threshold
of 1 gross domestic product (GDP) per capita and 3 GDP per capita. The intervention
is considered “cost-effective” if the ICER is below 1 GDP per capita and as “potentially
cost-effective” if it is less than 3 times GDP per capita
Sensitivity Analysis
A deterministic and a probabilistic sensitivity analyses were performed, the latter
using Monte Carlo simulations.
Results
Model for Caesarean Section
The cost-effectiveness ratio of carbetocin compared with oxytocin is presented in
[Table 4 ]. Carbetocin is the most effective treatment, and it is less expensive compared with
oxytocin, making it the dominant alternative. When considering the outcome of PPH
incidence, it is observed that for each hemorrhagic event avoided with the use of
carbetocin Col$ 94,887 are saved.
Table 4
Results of the base case
Cost-utility
QALY
Cost
ΔQALY
Δ Cost
ICER ($/QALY)
Carbetocin
0.9959
$ 461.750
0.00337
($ 20,116)
Dominant
Oxytocin
0.9926
$ 481.866
Cost-effectiveness
Incidence PPH
Cost
Avoided events
Δ Cost
ICER ($/Avoided PPH)
Carbetocin
0.230
$ 461,750
0.212
($ 20,116)
($ 94,887)
Oxytocin
0.442
$ 481,866
According to the deterministic sensitivity analysis, at the proposed intervals, the
variable that can impact the ICER, so that it is below the threshold, is the proportion
of patients presenting with hemorrhagic events with each of the alternatives. Since
this variable corresponds to a measure of drug effectiveness, it will not vary significantly
until new and more robust evidence is available.
A probabilistic sensitivity analysis was performed with the probability distributions
related to the methodology for each of the variables and a Monte Carlo simulation
was performed with 1,000 iterations to evaluate the consistency of the results ([Fig. 3 ]). A willingness to pay threshold of Col$ 53,090.188 (3 times the per capita GDP
of Col$ 17,696.729) was established. In all cases, carbetocin was taken as baseline
and oxytocin as a comparator. In 52% of the iterations carbetocin is maintained as
a dominant alternative since it is more effective and less costly. In 16.4% of the
iterations carbetocin is dominated by oxytocin. In 16.3%, it is more expensive and
more effective, but it is above the willingness to pay threshold, while in 15.3%,
it is below the threshold.
Fig. 3 Monte Carlo simulations in the cost-effectiveness incremental plane.
The acceptability curve ([Fig. 4 ]) shows that carbetocin proves to be a cost-effective alternative in 70% of the simulations
considering a threshold of 3 GDP/per capita (Col$ 53,090.188).
Fig. 4 Acceptability curves.
Model for Vaginal Delivery
The cost-effectiveness ratio of carbetocin compared with oxytocin is presented in
[Table 5 ]. Carbetocin is the most effective treatment, but it is also costlier compared with
oxytocin. The difference in terms of QALYs and incidence of PPH is small between the
two treatments. The ICER is above the cost-effectiveness threshold adopted by Colombia,
and when considering the outcome of PPH incidence it is observed that the cost required
to avoid additional hemorrhagic events is high.
Table 5
Results of the base case
Cost-utility
QALY
Cost
Δ QALY
Δ Cost
ICER ($/QALY)
Carbetocin
0.99800
$ 347,750
0.00009
$ 85,259
$974,790.719
Oxytocin
0.99791
$ 262,491
Cost-effectiveness
Incidence PPH
Cost
Avoided events
Δ Cost
ICER ($/Avoided PPH)
Carbetocin
0.156
$ 347,750
0.008
$ 85,259
$ 10,657.375
Oxytocin
0.164
$ 262,491
Abbreviations: ICER, incremental cost-effectiveness ratio; PPH, postpartum hemorrhage;
QUALY, quality-adjusted life year.
According to the univariate sensitivity analysis, the variables that can impact the
ICER are the proportion of patients presenting with bleeding events in each group
as well as the cost of carbetocin.
A probabilistic sensitivity analysis was performed considering the costs and other
variables that could become a source of heterogeneity. The probability distributions
related in the methodology for each of the variables were taken into account, and
a Monte Carlo simulation was performed with 1,000 iterations to evaluate the consistency
of the results. The threshold, again, was Col$ 53,090.188. [Fig. 5 ] presents the Monte Carlo simulation for incremental analysis. In all cases, carbetocin
was taken as the comparator and oxytocin as the baseline. It is observed that in 46.8%
of the iterations, oxytocin remains a dominant alternative since it is more effective
and less costly. In 27% of iterations carbetocin is the dominant alternative. In 18.9%
of the iterations carbetocin is more effective and more expensive than oxytocin but
is above the willingness to pay, while at 7.3% it is below the willingness to pay.
Fig. 5 Monte Carlo simulations in the cost-effectiveness incremental plane.
According to the acceptability curve, only carbetocin appears as a cost-effective
alternative in 30% of the simulations considering a threshold of 3 times per capita
GDP but it does not substantially increase its likelihood of being cost-effective,
even with a willingness to pay of Col$ 200,000.000.
Discussion
The results obtained are consistent with those found in other studies. According to
the evaluation developed by Moosivand et al,[14 ] from the perspective of Iran's health system and society, carbetocin is a dominant
alternative compared with oxytocin, considering the best possible scenarios (maximum
oxytocin price); however, is not cost-effective in other scenarios. In the probabilistic
sensitivity analysis, 51.7% of the iterations were below the willingness to pay threshold
for QALY.
In the model of women with at least one risk factor for PPH during vaginal delivery,
carbetocin is more effective than oxytocin; however, from the point of view of health
system and willingness to pay in the country, it is not a cost-effective alternative.
The difference in costs between alternatives is close to Col$ 80,000 but the difference
in terms of QALY is small. According to the literature review, no studies were identified
in vaginal delivery, which may be related in part to the fact that carbetocin does
not appear as a recommendation of choice in the prevention of PPH due to uterine atony
after vaginal delivery.[1 ]
[4 ]
[5 ]
[6 ]
[9 ]
The limitations of this study are related to the assumptions applied and the lack
of studies performed in the high-risk population, both in patients undergoing vaginal
delivery and cesarean delivery. The risk of hemorrhage and the use of additional uterotonics
differs from one study to the other, in part because of different oxytocin dosing.
We used the dosing scheme (and the PPH risk) described by Borruto et al[22 ] and Attilakos et al[23 ] because they included women with at least one risk factor for PPH. In other studies,
using even higher doses of oxytocin,[26 ]
[27 ]
[28 ]
[29 ] the risk reduction with carbetocin was even greater; however, there are differences
in the characteristics of the population, particularly related to the level of risk
that does not allow to clearly observe the differences between alternatives, so it
was preferred to adopt a conservative scenario. Regarding utility values, these were
based on international studies, since there are no quality of life studies available
in our country, and these may not necessarily reflect the preferences of Colombian
patients. In addition, in certain cases, the utility levels referred to in the literature
were obtained from a group of experts, so the estimated values may not correspond
to real life.
The costing of the different health states and events of the model was performed with
the help of clinical experts through the development of a typical case, since there
are restrictions to access databases with this type of information and considering
that the group of experts is small the estimated resources may not necessarily correspond
to those used in the country.
Conclusion
In this study, the economic model developed for cesarean section shows that carbetocin
has greater effectiveness and lower costs, making it a dominant alternative for the
prevention of PPH in high-risk patients. Sensitivity analyses show that within the
price range reported in SISMED, it remains either as a dominant alternative, or is
below the willingness to pay threshold. The economic model developed for vaginal delivery
shows that carbetocin has a cost per additional QALY higher than the acceptable threshold
for Colombia.