University of Groningen Pharmacological approaches to optimize TB treatment Zuur, Marlies

(1)

University of Groningen

Pharmacological approaches to optimize TB treatment

Zuur, Marlies

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2018

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Zuur, M. (2018). Pharmacological approaches to optimize TB treatment: An individualized approach. Rijksuniversiteit Groningen.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

Chapter

(3)

International Journal of

Tuberculosis and Lung Disease

2018;22(9):991-999

Cost-utility

Analysis of Treating

Tuberculosis Patients

with Intermediate

Susceptibility

dose-dependent with a

Higher Rifampicin

and Isoniazid

Dose in Europe

Marlanka A. Zuur,

Thea van Asselt,

Natasha van ‘t

Boveneind-Vrubleuskaya,

Alena Aleksa,

Maarten J. Postma,

Jan-Willem C. Alffenaar

(4)

Abstract

Setting

We propose to introduce a dose-dependent intermediate susceptibility category for Mycobacterium tuberculosis. Additionally, we propose to treat patients with

Mycobacterium tuberculosis in this category with a high dose of rifampicin and

isoniazid.

Objective

The aim of this study is to examine the impact of our strategy on Quality Adjusted Life Years and costs for both a country with a high multidrug-resistant tuberculosis prevalence and a low income (Belarus) and a country with a low multidrug-resistant tuberculosis prevalence and a high income (the Netherlands).

Design

A Markov model, consisting of 14 health states, was used to simulate the treatment outcomes and costs accrued over 5 years for a hypothetical cohort of 10,000 patients. One-way sensitivity analysis, probabilistic sensitivity analysis and a scenario analysis were also performed.

Results

Our strategy was shown to be cost-effective for Belarus and not cost-effective for the Netherlands. At a willingness-to-pay of 50,000 per Quality Adjusted Life Year, the probability of our strategy being cost-effective is 50% for the Netherlands and 57% for Belarus.

Conclusion

This study shows that our strategy could be cost saving and more effective. However, more studies are needed on the outcomes of using higher doses of isoniazid and rifampicin.

(5)

Introduction

Appropriate selection of critical concentrations used in drug susceptibility testing (DST) to determine whether the Mycobacterium tuberculosis

(Mtb) strain is resistant, is of great importance [1]. Based on pharmacokinetic

(PK) and pharmacodynamic (PD) considerations, it is postulated that the minimal inhibitory concentration (MIC) cut-off values above which therapy fails are significantly lower than the currently used breakpoints [1]. As a result of these lower critical concentrations, more patients would be classified as having multidrug-resistant tuberculosis (MDR-TB) and treated accordingly with more toxic and more expensive drugs [2]. Alternatively, an increased dose of isoniazid and rifampicin could be an attractive solution resulting in the same PK/PD indices, without increasing the number of MDR-TB patients [3]. Traditionally, the susceptibility of Mtb is classified as susceptible or resistant. By introducing dose-dependent intermediate susceptibility (DDIS), as has been done for many other bacteria, the gap between the currently used critical concentrations and the postulated new critical concentrations can be bridged [4]. Therefore, a second DST at the new concentrations needs to be performed for all Mtb strains that are drug susceptible according to current criteria [1]. Although, this strategy could lead to an increase in costs, it may have substantial benefit for TB patients. Increasing the dose may even be less expensive in the long-term, because more patients will have sufficient drug exposures resulting in less treatment failure, relapse and ADR [5]. A dose up to 35 mg/kg rifampicin and 16-18 mg/kg isoniazid is considered to be tolerable[6,7].

In this study we aimed to determine the impact of this proposed strategy by assessing the cost-effectiveness of performing a second Mycobacterium Growth Indicator Tube (MGIT) DST in combination with a higher dose of isoniazid and rifampicin in case of DDIS in different settings with respect to MDR-TB prevalence and income.

Materials and methods

Model approach

A Markov model was used to evaluate the costs and effects of performing a second DST and subsequent treatment with high dose isoniazid and rifampicin, as compared to current care, i.e. performing one DST. The model simulated the treatment outcomes and costs accrued over 5 years for a hypothetical cohort of 10,000 patients. This cohort started in the untested TB health state [Figure 1], where after the cohort distributed over various health states, known as Markov states. The patients moved

(6)

through the Markov model [Figure 1] based on transition probabilities [Table 1A].

If a patient relapsed or failed on treatment, the patient could enter any of the health states except for the untested health state. Patients could only enter the high-dose treatment state (shown in black in Figure 1), when a second MGIT DST was performed. This does not mean that intermediate susceptible TB could be ruled out, it was merely not confirmed. The incremental cost-effectiveness ratio (ICER) was calculated by dividing the incremental costs, i.e. the difference between the total costs of performing either strategy, with the incremental effectiveness of each strategy. The cost-effectiveness of this strategy was assessed for two European countries: Belarus and the Netherlands, which respectively are a low- and a high-income country as well as a high- and low MDR-TB prevalence country. The base case scenario considered TB patients of 18 years and older, because there were no studies available on the use of high-dose rifampicin and isoniazid in children. Future

Figure 1: General structure of the Markov model.

Patients that have an unknown tuberculosis infection enter the model. The black part of the model can only be entered by introduction of the second Mycobacteria Growth Indicator Tube drug-susceptibility test, otherwise patients can only be treated with normal-dose first-line anti-tuberculosis drugs or second-line anti-tuberculosis drugs. Death can be entered from any stage. UTB: Untested tuberculosis infection; NTB: normal-susceptible tuberculosis; MDRTB: Multidrug-resistant tuberculosis; NST: Normal-susceptible anti-TB treatment; HDT: high-dose first-line anti-TB treatment; SLT: second-line treatment; NSCT: Normal-susceptibility completed treatment; NSTF: Treatment failure after normal-susceptibility treatment; HDCT: Completed high-dose first-line anti-tuberculosis treatment; HDTF: Treatment failure after high-dose first-line tuberculosis drugs; SLCT: Completed second-line anti-tuberculosis treatment; SLTF: Treatment failure after second-line anti-anti-tuberculosis drugs.

(7)

costs and effects were discounted at a rate of 4% and 1.5% respectively, following Dutch guidelines in the absence of Belarusian guidelines [9].

Model variables

Transition probabilities of moving between health states, utilities and treatment costs are shown in Table 1. Cycle time was one month. Most probabilities were monthly probabilities and some were probabilities that only occurred once, such as the probability of having drug-susceptible-, DDIS- or MDR-TB, treatment success and ADR [Table 1A].

TB treatment was given for 6 months for dru-susceptible- and DDIS-TB and 20 months for MDR-TB [10,11]. A patient could fail on treatment at any given time, however treatment success could only occur after the above described treatment duration. The probability of treatment success, treatment failure and death for drug-susceptible as well as MDR-TB were obtained from the 2016 report on TB surveillance and monitoring in Europe [12]. Because these were yearly rates, the rates were converted to monthly probabilities according to the guideline for conducting economic evaluations in health care[9]. Relapse rates are shown in this report as relapse rates for drug-susceptible TB and MDR-TB together. The probability of relapse was based on independent studies that showed that the probability of relapse was four times higher for MDR-TB [11,13,14]. The percentage of patients with DDIS was based on the distribution of wild-type strains within this susceptibility range from three studies [15-17]. We included a 1.5 times higher treatment failure probability on normal dose treatment compared to high dose treatment based on a study from South Africa and Tanzania [18]. A study by Hu and colleagues showed that high-dose rifampicin and isoniazid could lead to around 50% less ADR and relapse than normal-dose treatment [19]. Therefore, we assumed ADR and relapse for high-dose treatment to be 50% less than for normal-dose treatment. The probability of death during high-dose treatment was estimated to be the same as normal-dose treatment. The probability of death during treatment was not based on all-cause mortality, but solely on sickness-related mortality i.e. TB.

Health outcomes

QALYs were calculated by multiplying the proportion of patients in a certain health state by the utility associated with that particular health state. Utility weights for the treatment of normal-susceptible TB (0.69), the treatment of MDR-TB (0.51) and for after the treatment of TB (0.88) were derived from the study by Kittikraisak et al. [20]. This study used the EuroQoL-5D to generate utility weights [21].

(8)

Table 1: Parameters used in the cost-effectiveness analysis Table 1A: Probabilities and overall mortality rates

Country

Transition probabilities (per month)

SE

Distri-bution used in PSA

References Untreated TB to drug-susceptible TB The Netherlands 0.991 y Fixed Fixed [12] Belarus 0.655 y Fixed Fixed [12] Undefined TB to multidrug-resistant TB The Netherlands 0.009 y 1.07*10 -3 Beta [12] Belarus 0.345 y 1.71*10 -4 Beta [12]

Drug-susceptible TB to dose-dependent intermediate susceptible TB

0.845

y

Fixed

Estimation

Normal-dose first-line anti-TB treatment to treatment failure

The Netherlands 6.69*10 -4 Fixed Fixed Estimation Belarus 3.58*10 -3 2.58*10 -5 Beta [12]

High-dose first-line anti-TB treatment to treatment failure

The Netherlands 4.42*10 -4 Fixed Fixed Estimation Belarus 2.36*10 -3 Fixed Fixed Estimation

Second-line anti-TB treatment to treatment failure

The Netherlands 3.76*10 -3 3.75*10 -4 Beta [31] Belarus 9.37*10 -3 3.45*10 -5 Beta [12]

Normal-dose first-line anti-TB treatment to treatment success

The Netherlands 87% y 1.58*10 -4 Beta [12] Belarus 86% y 4.45*10 -5 Beta [12]

High-dose first-line anti-TB treatment to treatment success

89% y 1.73*10 -3 Beta [18]

Second-line anti-TB treatment to treatment success

The Netherlands 73% y 2.20*10 -2 Beta [12] Belarus 54% y 1.81*10 -4 Beta [12]

Treatment success to death

(9)

Country Transition probabilities (per month)

SE

References

Normal-dose first-line anti-TB treatment to death

High-dose first-line anti-TB treatment to death

Second-line anti-TB treatment to death

The Netherlands 2.47*10 3.52*10 -4 Beta [31] Belarus 2.02*10 -2 3.63*10 -5 Beta [12] Untreated TB to death The Netherlands 1.12*10-2 7.51*10-7 Beta [33] Belarus 2.28*10-2 1.13*10-6 Beta [33]

Normal-dose first-line anti-TB treatment failure to MDR-TB

0.722

y

1.43*10-2

Beta

[34]

High-dose first-line anti-TB treatment failure to MDR-TB

0.361 y Fixed Fixed Estimation Relapse to MDR-TB The Netherlands 1.11*10 -2 5.47*10 -3 Beta [12] Belarus 9.33*10 -2 1.45*10 -4 Beta [12]

Normal-dose first-line anti-TB treatment success to relapse

High-dose first-line anti-TB treatment success to relapse

The Netherlands 2.50*10 -5 Fixed Fixed Estimation Belarus 2.09 *10 -3 Fixed Fixed Estimation

Second-line anti-TB treatment success to relapse

The Netherlands 2.00*10 -4 Fixed Fixed Estimation Belarus 1.80*10 -2 Fixed Fixed Estimation Table 1A continued

3b

(10)

Table 1B Costs

Country

Costs (Euro)

SE

References MGIT The Netherlands 37 Fixed Fixed [9] Belarus 7 Actual costs

Pharmacy costs drug-susceptible

treatment

The Netherlands

94

[9]

Normal-susceptible TB drug costs first

2 months The Netherlands 109 [9] Belarus 9 Actual costs

Normal-susceptible TB drug costs after

the first 2 months

The Netherlands 52 [9] Belarus 19 Actual costs Dose-dependent intermediate susceptible TB drug costs first 2 months The Netherlands 212 [9] Belarus 38 Actual costs Dose-dependent intermediate

susceptible TB drug costs after the first

2 months The Netherlands 155 [9] Belarus 30 Actual costs

Pharmacy costs MDR-TB treatment

The Netherlands 87 [9] MDR-TB drugs The Netherlands 5470 [9] Belarus 88 Actual costs Hospitalization drug-susceptible TB The Netherlands 177 [9] Belarus 8 Actual costs Hospitalization MDR-TB The Netherlands 1448 [9] Belarus 83 Actual costs Intake by a nurse The Netherlands 169 [9]

Social support from a nurse

The Netherlands 42 [9] DOT The Netherlands 281 [9]

(11)

Table 1B continued

Country

Costs (Euro)

SE

References

Production days lost

normal-susceptible TB The Netherlands 435 Fixed Fixed [9] Belarus 144 Actual costs

Production days lost Intermediate- susceptibility dose-dependent

The Netherlands 435 [9] Belarus 144 Actual costs

Production days lost MDR-TB

The Netherlands 387 [9] Belarus 128 Actual costs Diagnosis drug-susceptible TB The Netherlands 350 [9] Belarus 19 Actual costs Diagnosis MDR-TB The Netherlands 862 [9] Belarus 19 Actual costs Follow-up drug-susceptible TB The Netherlands 95 [9] Belarus 13 Actual costs Follow-up MDR-TB The Netherlands 54 [9] Belarus 12 Actual costs

Traveling costs patient

The Netherlands 1 [9] Belarus 3 Actual costs

3b

(12)

Table 1C: Utilities

Utility weight

SE

References

First-line anti-TB treatment

0.69

1.51*10

-3

Beta

[20]

Post first-line anti-TB treatment

0.88

9.03*10

-4

Beta

[20]

Second-line anti-TB treatment

0.51 4.01*10 -3 Beta [20] Post second-line anti-TB treatment 0.88 9.03*10 -4 Beta [20] Relapse 0.69 1.51*10 -3 Beta Estimated

TB: Tuberculosis; PSA: Probabilistic sensitivity analysis; MDR-TB: Multidrug-resistant Tuberculosis; MGIT: Mycobacteria Growth Indicator Tube; DOT: Directly-observed Therapy.

(13)

The utility of relapse was given a utility weight of 0.69, which is equal to the utility weight of normal-susceptible TB, because it was thought to be the same as not having received any treatment. High-dose treatment was also assigned the utility weight of normal-susceptible TB, since isoniazid and rifampicin are well tolerated [22].

Costs

The economic analysis was conducted from the societal perspective. All costs are in Euros. Costs for Belarus were converted to Euros according to the exchange rates in 2016. Conversion was not performed with purchasing power parities because these were not available for Belarus. For direct medical costs, we included costs of TB and TB control in the Netherlands [23]. These costs are from 2009 and were therefore converted into costs in 2016 using price index numbers from the Central Bureau of Statics as advised by Dutch guidelines [24]. Hospitalization duration was shown to be an average of 13.3 days for susceptible TB and 108.9 days for MDR-TB [23]. DOT and social support from a nurse were not provided in Belarus and pharmacy costs were included in the drug prices. The costs associated with retreatment were also included, since we included the possibility of relapse within five years. Travel costs were also taken into account, using the average distance to a hospital and average transportation costs in Belarus and the Netherlands. We estimated an average of one visit per month based on the average amount of visits during treatment of drug-susceptible TB [23]. Productivity costs, occurring when patients could not perform paid work, were also included. The productivity costs were calculated according to the human capital approach, using the average income in Belarus and the Netherlands and the average number of working days lost. The average number of working days lost was shown to be 81 days for drug-susceptible TB and 240 days for MDR-TB [25-27].

Sensitivity analysis

One-way sensitivity analysis was performed to examine robustness of outcomes, varying input variables along their reasonable ranges. All estimated transition probabilities were varied. Transportation costs were calculated based on 0 or 2 visits per month. DDIS was varied between 80 and 89%, because this was the case for respectively rifampicin and isoniazid [15-17]. The results are shown in tornado diagrams [Figure 2]. The results can help make a rough estimation of the cost-effectiveness of our approach in other countries besides the Netherlands and Belarus, by looking at a country’s characteristics. A probabilistic sensitivity analysis (PSA) was performed by varying all parameters simultaneously within their

(14)

respective distributions [Table 1] over 1,000 Monte Carlo simulations, to determine joint uncertainty [Figure 3]. From the Monte Carlo simulations a cost-effectiveness acceptability curve (CEAC) was constructed for both countries, which shows the probability of the approach being cost-effective at a willingness-to-pay (WTP) of 50,000 Euros [Figure 4].

Scenario analysis

The parameters that showed to have a large influence on the results based on the one-way sensitivity analysis were used to perform a scenario analysis. With the scenario analysis the probabilities were varied over a range in order to investigate when the approach would be most likely to be cost-effective. The utility of high-dose treatment could be higher, because the chance of relapse and failure is lower [19]. The utility of high-dose treatment could also be lower, because patients could experience side effects [11].

Results

The results of the cost-utility analysis are shown in Table 2 for Belarus and the Netherlands separately. Our strategy of treating a patient with DDIS-TB with high-dose first-line anti-TB treatment was shown to be cost saving and more effective, i.e. dominant, for Belarus. Our strategy was shown to be more expensive in the Netherlands and the QALY gain could not weigh up to this, which means that the strategy is not cost-effective for the Netherlands.

As can be seen from the tornado diagram [Figure 2], the assumptions made for the utility of (post) high-dose anti-TB treatment and the probability of death during high-dose anti-TB treatment, had the largest impact on the ICER for both Belarus as well as the Netherlands. These three parameters had such a large impact on the ICER of Belarus, that the influence of the other parameters became almost negligible. In the Netherlands treatment success also had a large influence on the ICER, however for Belarus the effect of this parameter on the ICER was not as apparent. Based on these results our approach would only be not cost-effective in Belarus if the utilities of high-dose treatment were lower than normal-dose treatment. However, this is thought to be unlikely because high-dose rifampicin and isoniazid are well tolerated [6,11].

(15)

Table 2: Results base-case analysis Belarus and the Netherlands Country Costs (€) QALY Costs/QALY Undiscounted Discounted Undiscounted Discounted Undiscounted Discounted One MGIT DST The Netherlands 9137.36 8953.67 4.09 3.94 Belarus 3278.14 3111.79 3.30 3.18 Two MGIT DST The Netherlands 9410.33 9234.45 4.10 3.94 Belarus 3232.94 3076.50 3.32 3.20 Incremental The Netherlands 280.78 272.38 0.00 0.00 144610 136535 Belarus -35.29 -45.20 0.02 0.03 Dominant Dominant

MGIT: Mycobacterium Growth Indicator Tube; QALY: Quality Adjusted Life Year

(16)

-7000 -6000 -5000 -4000 -3000 -2000 -1000 0 1000 2000 3000 Incremental cost-effectiveness ratio

HDT deathrate (min 25% more than NDT; max 25% less than NDT) Utility HDT (min 0.57; max 0.77) Utiliy post HDT (min 0.67; max 1.00

HDT success (min 85%, max 95%)

ADR HDT (min 75% less than NDT; max 25% less than NDT Relapse rate (min 23%; max 26%) Failure HDT (min 25% less than NDT; max 45% less than NDT) Transportation costs (min 0 visits; max 2 visits) Utility relapse (min 0.57; max 0.77) Intermediate susceptibility (min 0.8; max 0.9)

B.

Figure 2: Tornado diagrams based on one-way sensitivity analysis.

A. The Netherlands and B. Belarus. The parameter and the ranges are shown on the left. The dark grey bar is the incremental cost-effectiveness ratio associated with the minimum value and the light grey bar is the incremental cost-effectiveness ratio associated with the maximum value.

ADR: Acquired drug resistance; HDT: High-dose line anti-TB treatment; NDT: Normal-dose first-line anti-TB treatment

-25000 25000 75000 125000 175000 225000 275000 325000

Incremental cost-effectiveness ratio

HDT deathrate (min 25% more than NDT; max 25% less than NDT) Utility post HDT (min 0.67; max 1.00)

Utility HDT (min 0.57; max 0.77) HDT success (min 85%, max 95%)

ADR HDT (min 75% less than NDT; max 25% less than NDT Relapse rate (min 0%; max 87%)

Failure HDT (min 25% less than NDT; max 45% less than NDT) Intermediate susceptibility (min 0.8; max 0.9) Utility relapse (min 0.57; max 0.77) Transportation costs (min 0 visits; max 2 visits)

(17)

As can be seen from Figure 2A, the high and low values of the parameters that had the most influence on the analysis, are on the same side of the point estimate. This was caused by a loss in QALYs, which lead to a negative ICER. This means that our approach is dominated, as it is more expensive and less effective.

The PSA showed that our approach remained more expensive in the Netherlands, however in around 50% of the cases QALYs were gained [Figure 3A]. The CEAC showed that at a WTP of 50,000 per QALY, the probability that the strategy of introducing DDIS is cost-effective is 50% for the Netherlands [Figure 4].

The approach was cost saving in almost all simulations for Belarus, as can be seen in Figure 3B. However, the probability of the approach being cost-effective for Belarus is 57%, since QALY’s were lost in 43% of the cases. Even though this was compensated by cost savings, society and decision makers are reluctant to adopt strategies that appear less effective than current care.

As can be seen in Figure 5, the probability of our approach being cost-effective for the Netherlands increased when the utility for high-dose treatment was higher than the utility for normal-dose treatment. Additionally, the probability of death during high-dose treatment needs to be lower than the probability of death on normal-dose treatment to be cost-effective. For Belarus the approach remained cost-effective.

(18)

Figure 3: Probabilistic Sensitivity Analysis with n=1000 Monte Carlo analysis.

A. The Netherlands B. Belarus.

A. -400 -200 0 200 400 600 -2 -1 0 1 2 Co st s Effects B. -100 -80 -60 -40 -20 0 20 40 60 80 100 -2 -1 0 1 2 Co st s Effects

(19)

Figure 5: Scenario analysis of a difference in the utility and death probability for high-dose first-line anti-tuberculosis treatment in Belarus and the Netherlands.

The utility rate ranges from 0.4 to 0.8 and the death rate from 40% less to 40% more chance of death than normal-dose first-line ant-TB treatment, with steps of 20%.

-300 -200 -100 0 100 200 300 -0,15 -0,1 -0,05 0 0,05 0,1 0,15 In cr em en ta l c os ts Incremental QALYs

Death rate high-dose treatment in relation to normal-dose treatment in the Netherlands Utility high-dose treatment the Netherlands

Death rate high-dose treatment in relation to normal-dose treatment in Belarus Utility high-dose treatment Belarus 40% less 40% less 0.4 0.5 0.6 0.7 0.8 40% more 0.4 0.5 0.6 0.7 0.8 40% more

Figure 4. Cost-effectiveness acceptability curve (CEAC)

A. The Netherlands B. Belarus

(20)

Discussion

To our knowledge, this is the first study showing the cost-effectiveness of the introduction of the DDIS category for TB and subsequent treatment with high-dose isoniazid and rifampicin. Only a few studies have been published on the cost-effectiveness of MDR-TB treatment. One study, by Dowdy et al., was similar to this study. However they used Disability Adjusted Life Years (DALYs) and not normal all-cause mortality before and after TB treatment [28]. This is mainly due to the fact that they used a time-horizon that only included the treatment duration [28]. The PSA showed that the probability of the approach being cost-effective is 57% for Belarus and 50% for the Netherlands at a WTP of 50,000 Euros. The results of the PSA reflected an uncertainty, showing that both strategies could still not be cost-effective. This uncertainty is thought to be mainly caused by the utility weights, which were shown to have a large influence on the results. This is because our strategy became not cost-effective when QALYs were lost. The utility weights that were used in this study were median values and not mean values. However, these two values are not expected to differ a lot, because the interquartile ranges were rather well distributed around the median.

Since TB treatment is already quite optimal in the Netherlands, high-dose treatment might not give as much additional benefit. Therefore the higher costs of the increase in dosage lead to an increase in costs, such that it could not be offset by the savings from improved failure and relapse rates. Our approach was shown to save costs in Belarus, because there was a much higher relapse rate in Belarus that resulted in a longer treatment period. The relapse rate could be reduced by 50% with high-dose treatment, therewith reducing the costs. Another difference between the Netherlands and Belarus is the background death rate which was almost twice as high in Belarus [29].

Limitations

Our model has some limitations. First, there was limited data on treatment outcomes of high-dose first-line anti-TB treatment, therefore assumptions were made. Most of the evidence serving as input for the model was derived from a recent study on high-dose rifampicin in a small group of patients [18]. Because the relapse rate and the amount of ADR for high-dose treatment could not be determined in this study, a probability of 50% less relapse than for normal-dose was used based on the in

vitro and in vivo study from Hu and colleagues [19]. Second, utility decrements for

(21)

impact on the quality of life for all adverse events during TB treatment. However, the risk of adverse events was shown to be twice as high for high-dose rifampicin and there could be costs related to this [11]. Costs for hospitalization and production losses were included in the analysis, which will in part comprise of the costs related to adverse effects, but our approach to not use a different utility might be conservative. Third, mono-resistance was not included, because this would make the model more complex. High-dose isoniazid is used for RR-TB and therefore, inclusion of mono-resistance could have a positive effect on the cost-effectiveness of our strategy [8,30].

Strengths

We took into account all treatment outcomes for all treatment strategies including the risk of death, ADR, treatment completion and treatment failure, as well as the costs of completing the different treatment strategies. Our model is comprehensive since it also incorporated relapse, by using a broader time horizon.

Conclusion

In conclusion, this study shows that costs and QALYs could be saved with the introduction of the DDIS-TB category and subsequent treatment with high-dose isoniazid and rifampicin in high MDR-TB prevalence and low-income countries, such as Belarus. In high-income countries with a relatively low MDR-TB prevalence, testing for DDIS-TB was shown to be not cost-effective, most likely because there is not enough room for improvement in TB treatment in these countries and the expensive treatment. More studies need to be performed to determine the effectiveness of high-dose first-line anti-TB treatment in larger cohorts, in order to take away some of the uncertainty in our results.

(22)

References

1. Gumbo T. New susceptibility breakpoints for first-line antituberculosis drugs based on antimicrobial pharmacokinetic/pharmacodynamic science and population pharmacokinetic variability. Antimicro.b Agents Chemother. 2010; 54: 1484-1491.

2. Marks SM, Flood J, Seaworth B, Hirsch-Moverman Y, Armstrong L, Mase S, Salcedo K, Oh P, Graviss EA, Colson PW, Armitige L, Revuelta M, Sheeran K, TB Epidemiologic Studies Consortium. Treatment practices, outcomes, and costs of multidrug-resistant and extensively drug-resistant tuberculosis, United States, 2005-2007. Emerg. Infect. Dis. 2014; 20: 812-821.

3. Alffenaar JW, Akkerman OW, Bolhuis MS, Boeree MJ, de Lange WC, van der Werf TS. Breakpoints and drug exposure are inevitably closely linked.

Antimicrob. Agents Chemother. 2015; 59: 1384-14.

4. Turnidge J, Paterson DL. Setting and Revising Antibacterial Susceptibility Breakpoints. Clin. Microbiol. Rev. 2007; 20: 391-408.

5. Pasipanodya JG, McIlleron H, Burger A, Wash PA, Smith P, Gumbo T. Serum drug concentrations predictive of pulmonary tuberculosis outcomes. J. Infect

.Dis. 2013; 208: 1464-1473.

6. Boeree MJ, Diacon AH, Dawson R, Narunsky K, du Bois J, Venter A, Phillips PP, Gillespie SH, McHugh TD, Hoelscher M, Heinrich N, Rehal S, van Soolingen D, van Ingen J, Magis-Escurra C, Burger D, Plemper van Balen G, Aarnoutse RE, PanACEA Consortium. A dose-ranging trial to optimize the dose of rifampin in the treatment of tuberculosis. Am. J. Respir. Crit. Care

Med. 2015; 191: 1058-1065.

7. Katiyar SK, Bihari S, Prakash S, Mamtani M, Kulkarni H. A randomised controlled trial of high-dose isoniazid adjuvant therapy for multidrug-resistant tuberculosis. Int. J. Tuberc. Lung Dis. 2008; 12: 139-145.

8. World Health Organization. WHO treatment guidelines for drug-resistant tuberculosis, 2016 update. 2016; WHO/HTM/TB/2016.04.

9. Zorginstituut Nederland. Guideline for conducting economic evaluations in health care. 2016.

10. World Health Organization. Treatment of Tuberculosis Guidelines. World Health Organization, 2010.

(23)

11. Steingart KR, Jotblad S, Robsky K, Deck D, Hopewell PC, Huang D, Nahid P. Higher-dose rifampin for the treatment of pulmonary tuberculosis: a systematic review. Int. J. Tuberc. Lung Dis. 2011; 15: 305-316.

12. European Centre for Disease Prevention and Control/WHO Regional Office for Europe. Tuberculosis surveillance and monitoring in Europe 2016. Stockholm: European Centre for Disease Prevention and Control 2016. 13. Millet JP, Shaw E, Orcau A, Casals M, Miro JM, Cayla JA, Barcelona

Tuberculosis Recurrence Working Group. Tuberculosis recurrence after completion treatment in a European city: reinfection or relapse? PLoS One. 2013; 8: e64898.

14. Becerra MC, Appleton SC, Franke MF, Chalco K, Bayona J, Murray MB, Mitnick CD. Recurrence after treatment for pulmonary multidrug-resistant tuberculosis. Clin. Infect. Dis. 2010; 51: 709-711.

15. Schon T, Jureen P, Giske CG, Chryssanthou E, Sturegard E, Werngren J, Kahlmeter G, Hoffner SE, Angeby KA. Evaluation of wild-type MIC distributions as a tool for determination of clinical breakpoints for

Mycobacterium tuberculosis. J. Antimicrob. Chemother. 2009; 64: 786-793. 16. Wanger A, Mills K. Testing of Mycobacterium tuberculosis susceptibility to

ethambutol, isoniazid, rifampin, and streptomycin by using Etest. J. Clin.

Microbiol. 1996; 34: 1672-1676.

17. Gumbo T, Pasipanodya JG, Wash P, Burger A, McIlleron H. Redefining multidrug-resistant tuberculosis based on clinical response to combination therapy. Antimicrob. Agents Chemother. 2014; 58: 6111-6115.

18. Boeree MJ, Heinrich N, Aarnoutse R, Diacon AH, Dawson R, Rehal S, Kibiki GS, Churchyard G, Sanne I, Ntinginya NE, Minja LT, Hunt RD, Charalambous S, Hanekom M, Semvua HH, Mpagama SG, Manyama C, Mtafya B, Reither K, Wallis RS, Venter A, Narunsky K, Mekota A, Henne S, Colbers A, van Balen GP, Gillespie SH, Phillips PPJ, Hoelscher M; PanACEA consortium. High-dose rifampicin, moxifloxacin, and SQ109 for treating tuberculosis: a multi-arm, multi-stage randomised controlled trial. Lancet Infect. Dis. 2017; 17(1):39-49.

19. Hu Y, Liu A, Ortega-Muro F, Alameda-Martin L, Mitchison D, Coates A.

(24)

dose rifampicin kills persisters, shortens treatment duration, and reduces relapse rate in vitro and in vivo. Front. Microbiol. 2015; 6: 641.

20. Kittikraisak W, Kingkaew P, Teerawattananon Y, Yothasamut J, Natesuwan S, Manosuthi W, Chongsuvivatwong V, Whitehead SJ. Health related quality of life among patients with tuberculosis and HIV in Thailand. PLoS One. 2012; 7: e29775.

21. EuroQol Group. EuroQol--a new facility for the measurement of health-related quality of life. Health Policy 1990; 16: 199-208.

22. Goutelle S, Bourguignon L, Maire P, Jelliffe RW, Neely MN. The Case for Using Higher Doses of First Line Anti-Tuberculosis Drugs to Optimize Efficacy. Curr. Pharm. Des. 2014; 20: 6191-6206.

23. de Vries G, Baltussen R. Costs for tuberculosis and tuberculosis control in the Netherlands. Infectieziekten Bulletin. 2013; 24: 136-140.

24. Central Bureau of Statistics. Statline. Consumerprices; price index 2015=100. http://statline.cbs.nl/Statweb/publication/. june 2017. Date last accessed: June 22 2017.

25. Diel R, Nienhaus A, Lampenius N, Rusch-Gerdes S, Richter E. Cost of multi drug resistance tuberculosis in Germany. Respir. Med. 2014; 108: 1677-1687. 26. Kik SV, Olthof SP, de Vries JT, Menzies D, Kincler N, van

Loenhout-Rooyakkers J, Burdo C, Verver S. Direct and indirect costs of tuberculosis among immigrant patients in the Netherlands. BMC Public Health. 2009; 9: 283-2458-9-283.

27. National statistical committee of the Republic of Belarus. Belarus average monthly wages. https://tradingeconomics.com/belarus/wages. 2017. Date last accessed: July 31 2016.

28. Dowdy DW, van’t Hoog A, Shah M, Cobelens F. Cost-effectiveness of rapid susceptibility testing against second-line drugs for tuberculosis. Int. J. Tuberc.

Lung Dis. 2014; 18: 647-654.

29. World Health Organization. Better noncommunicable disease outcomes: challenges and opportunities for health systems: Belarus Country Assessment 2016.

30. World Health Organization. Global tuberculosis report 2016. WHO/HTM/ TB/2016.13

(25)

31. van Altena R, Akkerman OW, Alffenaar JC, Kerstjens HA, Magis-Escurra C,Boeree MJ, van Soolingen D, de Lange WC, Bolhuis MS, Hoefsloot W, de Vries G, van der Werf TS. Shorter treatment for multidrug-resistant tuberculosis: the good, the bad and the ugly.

Eur. Respir. J. 2016; 48: 1800-1802.

32. United Nations, Department of Economic and Social Affairs, Population Division. World Population Prospects: The 2015 Revision, Key Findings and Advance Tables. 2015; ESA/P/WP.241.

33. Tiemersma EW, van der Werf MJ, Borgdorff MW, Williams BG, Nagelkerke NJ. Natural history of tuberculosis: duration and fatality of untreated pulmonary tuberculosis in HIV negative patients: a systematic review. PLoS

One. 2011; 6: e17601.

34. Satti H, McLaughlin MM, Seung KJ, Becerra MC, Keshavjee S. High risk of drug-resistant tuberculosis when first-line therapy fails in a high HIV prevalence setting. Int. J. Tuberc. Lung Dis. 2013; 17: 100-106.