Cost-effectiveness of ceftolozane/tazobactam compared to meropenem in ventilator-associated pneumonia

Abstract

Introduction

Ventilator-associated pneumonia (VAP) is a prominent cause of morbidity and mortality in intensive care unit patients. Ceftolozane/tazobactam could be a cost-effective treatment option prior to culture test result for treatment of VAP. This study aims to evaluate the cost-effectiveness of ceftolozane/tazobactam compared to meropenem in Colombia.

Methods

We constructed a cost-effectiveness model using a short-term decision tree, to evaluate the cost-effectiveness of C/T for the treatment of patients VAB versus meropenem. The model was analyzed probabilistically, and a value of information (VOI) analysis was conducted to inform the value of conducting further research to reduce current uncertainties in the evidence base. Cost-effectiveness was evaluated at a willingness-to-pay (WTP) value of US$4,487.

Results

Meropenem has a lower cost of US$ 6042, while CT has a higher cost of US$ 6324, with an incremental cost of US$ 281 compared to Meropenem. In terms of effectiveness, the probability of cure (C/T = 72%, Meropenem = 85%). CT shows a slightly higher effectiveness with an incremental effectiveness of 1.0%. The incremental cost-effectiveness ratio (ICER) is calculated at US$ 27,869 per cured patient. The base case results were robust to most of variations in utilities, transition probabilities, relative risk. The price threshold analysis for the cost for intravenous powder for injection of CT (1 g) shows if the cost per injection es lower than US$36; that it a reduction of 32% in it cost, this alternative is cost-effective compared to meropenem.

Conclusion

Ceftolozane/tazobactam is not a cost-effective option compared to Meropenem to treat patients with VAP in Colombia. If the cost for intravenous powder for injection of Ceftolozane/tazobactam is reduced by 32%, this alternative could be cost effective respect to meropenem.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12962-025-00658-0.

Introduction

Ventilator-associated pneumonia (VAP) is a prominent cause of morbidity and mortality in intensive care unit (ICU) patients, with mortality rates ranging from 16% to 94% in developing countries [1]. The prevalence of multidrug-resistant Gram-negative pathogens, particularly Pseudomonas aeruginosa imposes challenges to seek effective, safe, and efficient treatments [2]. Ceftolozane/tazobactam (CT), a combination of a novel cephalosporin and a β-lactamase inhibitor, has demonstrated efficacy in treating VAP and has a broad susceptibility profile against Gram-negative pathogens [3].

Inappropriate treatment of VAP leads to adverse outcomes and increased healthcare costs [4]. VAP are associated with high ICU costs, hospitalizations, and lost productivity. A cohort study of 288,538 patients confirms that patients with VAP have higher hospital costs (median: $64,639, IQR: $36,979-$108,391) and longer hospital stays (median: 21 days, IQR: 13–33 days) [5]. Indeed, based on a model that synthesized evidence primarily from a Phase III trial, Naik et al. estimated that ceftolozane/tazobactam could be a cost-effective treatment option in both early (prior to culture test result) and confirmed treatment for VAP, compared to meropenem in US [6]. Ceftolozane/tazobactam has been suggested as empirical therapy due to its high activity against multidrug-resistant Pseudomonas aeruginosa and other Gram-negative bacilli commonly implicated in VAP in Colombia [7]. Given the diagnostic delays in obtaining microbiological confirmation in critical care, empirical use of broad-spectrum agents remains a frequent and necessary clinical strategy in many ICUs [8]. Although de-escalation based on culture is desirable, local data on pathogen resistance patterns often justify initial use of agents like C/T [9]. To our knowledge, no cost-effectiveness analysis in Colombia has yet addressed the targeted use of C/T based on culture results, reinforcing the relevance of our empirical-use scenario. Ceftolozane/tazobactam could not be a cost-effective option in developing countries; due to the lower willingness to pay per life year gained in developing countries, higher cost of new drugs, challenges in healthcare infrastructure, resource limitations, and access to advanced diagnostic [10]. This study aims to evaluate the cost-effectiveness of ceftolozane/tazobactam compared to meropenem, considering real-world treatment scenarios, clinical efficacy data, and susceptibility profiles, to inform decision-making and address the urgent need for effective therapies against Gram-negative infections.

This study was conducted in Colombia, a middle-income country facing the dual challenge of a high burden of ventilator-associated pneumonia and constrained access to new antimicrobials due to limited healthcare budgets. By incorporating local data on drug costs, bacterial susceptibility, and healthcare delivery, the findings are highly relevant to national policymakers, third-party payers, and formulary decision-makers, providing actionable evidence for resource allocation and pricing negotiations.

Materials and methods

We developed a cost-effectiveness analysis of CT for the treatment of patients VAB versus meropenem, the reference comparator of the ASPECT-NP pivotal trial [11]. We constructed a cost-effectiveness model using a short-term decision tree, published previously by Naik et al. The focus of the decision tree was on adults admitted to the ICU and receiving ventilation for VAP, consistent with the ASPECT-NP trial population. The ASPECT-NP trial enrolled adult patients admitted to the ICU with ventilated hospital-acquired or ventilator-associated bacterial pneumonia. The median age was 62 years, approximately 64% were male, and patients had high disease severity (mean APACHE II score of 17.2) [11]. Ceftolozane/tazobactam was administered at 3 g (2 g/1 g) IV every 8 h, and meropenem at 1 g IV every 8 h, for a typical duration of 8 to 14 days, according to the ASPECT-NP trial protocol. In our model patient response was classified as either ‘cured,’ ‘not cured,’ or ‘death (Fig. 1). Patients were initiated on either CT or meropenem prior to obtaining their susceptibility profile. We assumed a delay of approximately 2 days to receive confirmation of antimicrobial susceptibility or resistance from the date the culture was obtained, reflecting real-world delays in microbiological diagnostics in Colombia. Once the test results became available, patients followed one of three treatment pathways:

Fig. 1.

Decision tree model

1. If the causative pathogen(s) were susceptible to the initial treatment, they continued with that therapy.

2. If the patient did not achieve a cure with the initial treatment, subsequent treatment options were explored.

3. If the causative pathogen(s) were not susceptible to the initial treatment (considered as early inappropriate therapy), an alternative treatment was administered.

In patients with susceptibility and clinical response, a switch to a lower-cost alternative was also modeled, particularly relevant for CT, which incurs a higher acquisition cost than meropenem. This reflects local antimicrobial stewardship practices and third-payer cost containment strategies. The study was carried out from the perspective of the third payer. Only direct medical costs were included. The time horizon was 30 days, aligning with the ASPECT-NP trial and other economic evaluations of acute infectious diseases. Given this short-term horizon, discounting of costs and outcomes was not applied.

Inputs of the model

Effectiveness was derived from clinical cure and mortality rates reported in the ASPECT-NP trial and the Naik et al. model, see Supplementary material. Table 1s [6]. Relative risks were applied to model the differential mortality and clinical failure based on trial estimates. The distribution of subsequent treatments was determined based on susceptibility profiles obtained from a prospective multicenter study conducted across 8 cities in Colombia [12]. The clinical response, length of hospital stays (LOS), and all-cause mortality rates associated with CT and meropenem were extracted from a previously published model by Naik et al., which was built upon the ASPECT-NP trial [6]. The probability of de-escalation or switch treatment in patients with confirmed susceptibility was obtained from a multicenter study conducted in Latin America [13]. National surveillance data from Colombian ICUs indicate that only approximately 50% of P. aeruginosa isolates are susceptible to carbapenems such as meropenem, reinforcing the rationale for considering agents like ceftolozane/tazobactam in the empirical treatment of VAP in high-risk settings [14].

Costs were measured using a bottom-up approach from a previously published Colombian cost-of-illness study, which identified healthcare resource use through national guidelines and expert opinion using the nominal group technique. All costs were estimated over a 30-day period, consistent with the acute course of VAP. The cost inputs were derived from a cost-of-illness study on VAP previously conducted in Colombia. This study applied a bottom-up methodology identifying cost-generating events (direct health service consumption, diagnostic tests, drugs) based on national clinical guidelines. Quantities and frequencies of use were estimated using expert elicitation and nominal group techniques [15]. Cost were extracted from nationals list prices (ISS and SISMED) and in the sensitivity analysis a range between 25% and 50% variation was used to reflect uncertainty [16]. As ISS and SISMED tariffs tend to underestimate actual acquisition costs, a correction factor of + 30% was applied, consistent with previous economic evaluations in Colombia, to better approximate real-world costs [17–29]. We conducted face validity with clinical experts and pharmacists, structural validation by comparing model structure with similar models in the literature, and cross-validation of intermediate outcomes (e.g., LOS, cure rates) with published Colombian data and local clinical audits. All costs were adjusted for inflation to account for 2022 Colombian pesos. We used the average exchange rate of 2022 to transform Colombian pesos into U.S. dollars (currency rate: US$1.00 = COP$ 4,000) [30]. The incremental cost-effectiveness ratio (ICER) was calculated using the following formulae:

Also, we estimated the net monetary benefit (NMB). NMB represents the value of an intervention in monetary terms [31]. NMB is calculated as (incremental benefit x threshold) – incremental cost. Incremental NMB measures the difference in NMB between alternative interventions, a positive incremental NMB indicating that the intervention is cost-effective compared with the alternative at the given willingness-to-pay threshold.

Further analysis was conducted to determine the expected value of perfect information (EVPI) and the perfect parameter information (EVPPI) [32]. These calculations provide an upper bound for the value of future research to reduce current uncertainties, either for the entire model (EVPI) or for specific parameters (EVPPI). We used a nonparametric regression-based method and implemented the analysis using the Sheffield Accelerated Value of Information (SAVI) tool [33]. We assumed 1,000 VAP cases annually based on national hospital discharge data and ICU infection surveillance reports in Colombia [12]. The payer strategy-specific burden (PSB) and payer uncertainty burden (PUB) were calculated to reflect the payer’s financial risks using the method described by Grimm et al. [34]. Any treatment was considered cost-effective if the incremental cost-utility ratio was below US$ 4,487 following recent estimations of willingness to pay in Colombia [35].

Sensitivity analysis

Since probabilities did not come from the Colombian population, they were subjected to a probabilistic sensitivity analysis, as detailed below and as recommended by the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement [36]. This ensures transparency and robustness in economic evaluations, particularly in LMIC settings with high parameter uncertainty. For the one-way sensitivity analysis, we varied transition probabilities by ± 25% from their base-case value, reflecting plausible uncertainty in the absence of local data. For PSA, we used beta distributions for probabilities and utilities, and gamma distributions for costs, consistent with standard modeling practice. Full details of distributions and parameter ranges are provided in Supplementary Table S1. For each treatment strategy, we calculated the expected costs and benefits using the combination of all parameter values in the model. To represent decision uncertainty, we plotted the cost-effectiveness and acceptability frontiers. Microsoft Excel^® was used for all analysis. Ethics approval was not required, since the data used in this study came from publicly available data.

Results

Table 1 presents the results of a cost-effectiveness analysis comparing two strategies: Meropenem and CT. Meropenem has a lower cost of US$ 6042, while CT has a higher cost of US$ 6324, with an incremental cost of US$ 281 compared to Meropenem. In terms of effectiveness, the proportion of patients cured of Meropenem was 47% and CT was 48%. CT shows slightly higher effectiveness with an incremental effectiveness of 1.0%. The incremental cost-effectiveness ratio (ICER) is calculated to be US$ 27 869, indicating the additional cost required to achieve one unit of incremental effectiveness. This ICER is higher than willingness to pay in Colombia of $ 4487. The net monetary benefit (NMB) for Meropenem was -US$ 3640, while Ceftolozane/tazobactam shows a negative NMB of -US$ 3870.We estimated the incremental cost per death avoided. C/T was associated with a reduction in mortality (12.4% vs. 15.7%), corresponding to an absolute risk reduction of 3.3%. The incremental cost per death avoided was US$18,462, below the commonly accepted mortality WTP threshold in Colombia. These results suggest that Meropenem may be cost-effective compared to CT under current price conditions and susceptibility scenarios.

Table 1.

Base-case cost-effectiveness results comparing Ceftolozane/Tazobactam and meropenem for the treatment of ventilator-associated pneumonia in Colombia

Strategy	Cost (USD per patient)	Incremental Cost (US$)	Effectiveness(Proportion of patients cured)	Incremental Effectiveness	ICER(USD per additional cured patient)	NMB
Meropenem	6042		0,47			-3640
Ceftolozane/tazobactam	6324	282	0,48	0,01	27,870	-3870

NMB: Net monetary benefit (USD)

Sensitivity analysis

In the deterministic sensitivity analyses, our base case results were robust to most of variations in utilities, transition probabilities, relative risk, and cost; except for the length of stay (LOS) of patients cured by Meropenem and CT (see Fig. 2). It the length of stay of patients cured by meropenem is more than 23 days the incremental net monetary benefit is higher than 0 and, CT will be the cost-effectiveness treatment. Otherwise, the length of stay of patient cured by CT is less than 18.5 days the incremental net monetary benefit is higher than 0 and, CT will be the cost-effectiveness treatment.

Fig. 2.

Tornado diagram

Price threshold

The one-way sensitivity analysis of the cost for intravenous powder for injection of CT (1 g) shows if the cost per injection is lower than US$36, representing a reduction of approximately 32%—then CT becomes cost-effective compared to Meropenem (see Supplementary Fig. 1s).

The cost-effectiveness plane displays 10,000 Monte Carlo iterations with the majority falling in the northeast and southeast quadrants, suggesting increased cost and effectiveness with CT but variable cost-effectiveness. The CEAC curve shows the probability that CT is cost-effective across a range of WTP thresholds. At US$ 4,487, the probability is 0.389, favoring Meropenem, Fig. 3. The Cost-Effectiveness Acceptability Curve (CEAC) shows the probability that all strategies are cost-effective at varying thresholds. The results show that at a threshold value of US$ 4,487 per proportion of patients cured, the strategy with the highest probability of being most cost-effective is Meropenem, with a probability of 0.611, Fig. 4.

Fig. 3.

Cost effectiveness plane

Fig. 4.

Acceptability curve

Expected value of perfect information

The overall EVPI per person affected by the decision is estimated to be USD $5.82, indicating that additional research could have limited monetary value per patient. However, the population EVPI could be relevant if large patient cohorts are expected to benefit. The EVPPI highlighted that the highest uncertainty lies in the distribution of treatment efficacy and cost inputs, especially the LOS associated with each treatment arm.

Discussion

Our results indicate that Meropenem has a lower cost compared to CT, with a cost difference of US$ 281. However, CT demonstrates slightly higher effectiveness, with an incremental effectiveness of 1.0% compared to Meropenem. The incremental cost-effectiveness ratio (ICER) is calculated to be US$ 27 869, which exceeds the willingness to pay threshold in Colombia This suggests that, under current price and effectiveness assumptions, CT may not be a cost-effective intervention in this context.

As expected, hospital stay was a key driver in the model. If the length of stay of patients cured by meropenem is more than 23 days the incremental net monetary benefit is higher than 0 and CT will be the cost-effectiveness treatment. Otherwise, the length of stay of patient cured by CT is less than 18.5 days the incremental net monetary benefit is higher than 0 and, CT will be the cost-effectiveness treatment. This variable is highly sensitive to real-world effectiveness data and resource use patterns. Although it modifies the result in the deterministic sensitivity analysis, in most probabilistic simulations CT was not cost-effective compared to Meropenem.

There are higher differences in direct cost per injection between meropenem and CT; Meropenem´s cost per injection is 90% less than the cost per Ceftolozane/tazobactam´s cost per injection. CT is not cost effective at the WTP declared in Colombia. However, our one-way sensitivity analysis showed that if the cost of CT were reduced by approximately 32%, it could become a cost-effective alternative. This highlights one of our principal findings: the possibility of altering the cost-effectiveness status of CT by adjusting a single variable—its acquisition cost. The high cost of new antibiotics in developing countries represents a significant barrier to achieving better patient outcomes and exacerbates health inequities within these populations [37]. The limited access to affordable antibiotics leads to suboptimal treatment choices and compromises the ability to combat infectious diseases effectively [38]. This situation is further exacerbated by the fact that developing countries often face a higher burden of infectious diseases and drug-resistant pathogens [39]. The unaffordability of these antibiotics not only hinders the provision of adequate healthcare but also perpetuates disparities in health outcomes between developed and developing countries [40]. The high cost of new antibiotics in developing countries creates health economics problems by straining already limited healthcare budgets, diverting resources away from other essential health interventions, and impeding sustainable healthcare systems. Addressing this issue requires collaborative efforts from governments, international organizations, and pharmaceutical companies to ensure equitable access to affordable and effective antibiotics, promoting better health outcomes and reducing health inequity. Adverse events were not explicitly included in the model, as serious drug-related events were rare and comparable between ceftolozane/tazobactam and meropenem in the ASPECT-NP trial. Therefore, their omission is unlikely to meaningfully alter the results. Nonetheless, we acknowledge that minor differences in safety profiles could modestly impact total costs and outcomes.

This study has several limitations. First, the 30-day time horizon may underestimate long-term costs and outcomes. Second, our model is based on data from international sources due to limited availability of Colombian data, particularly for transition probabilities and utilities. While we applied local costing data and conducted extensive sensitivity analyses, the results should be interpreted cautiously and are best viewed as exploratory rather than definitive. Third, the ISS tariff schedule, while officially used, may not fully reflect current market prices despite the applied correction factor. Lastly, we applied a QALY-based willingness-to-pay threshold to a model based on cure proportions due to the absence of established thresholds for intermediate outcomes in Colombia. These limitations should be considered when interpreting the results. Our findings differ from previous evidence. Firstly, previous studies in US have suggested that CT may be a cost-effective option compared to Meropenem due to its higher effectiveness and favorable cost-effectiveness ratio [6]. In addition to the fact that in Colombia we have a WTP much lower than the WTP of the US, the main difference in the inputs between this study and our study, were the local data of costs of the drugs, and the susceptibility ratios to CT and meropenem. The other probabilities in the decision tree were the same as those published in that model due to the absence of local data. Even though the susceptibility values ​​extracted from local literature were lower than those used in the Naik model, the model and the results were robust to changes in these values. Our study has some limitations. We used a relative risk extracted from literature and not estimated directly from our population. However, the results of the probabilistic sensitivity analysis confirmed the robustness of the model’s results. The relative risk was subjected to a probabilistic sensitivity analysis, as detailed above and as recommended by the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) Statement [36].

In conclusion, although Ceftolozane/Tazobactam shows slightly higher clinical effectiveness, Meropenem appears to be more cost-effective in the Colombian context, given its substantially lower acquisition cost and marginal difference in health outcomes. Reducing the cost of Ceftolozane/Tazobactam could improve its cost-effectiveness profile. Our findings underscore the importance of local economic evaluations in guiding treatment decisions and antibiotic stewardship in resource-limited settings.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

WTP

Willingness to pay

CHEERS

Consolidated Health Economic Evaluation Reporting Standards

ICER

Incremental cost-effectiveness ratio

Author contributions

J.A.B and A.F.Z. wrote the main manuscript text and D.G. prepared Figs. 1, 2, 3 and 4. All authors reviewed the manuscript.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Institutional Review Board of University of Antioquia (2015–4690).

Financial disclosures

This study was supported by own funding of authors.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.