Cost-effectiveness of dipeptidyl peptidase-4 inhibitor monotherapy versus sulfonylurea monotherapy for people with type 2 diabetes and chronic kidney disease in Thailand

Abstract

Objective: With a high prevalence of chronic kidney disease (CKD) in type 2 diabetes (T2DM) in Thailand, the appropriate treatment for the patients has become a major concern. This study aimed to evaluate long-term cost-effective of dipeptidyl peptidase-4 (DPP-4) inhibitor monothearpy vs sulfonylurea (SFU) monotherapy in people with T2DM and CKD.

Methods: A validated IMS CORE Diabetes Model was used to estimate the long-term costs and outcomes. The efficacy parameters were identified and synthesized using a systematic review and meta-analysis. Baseline characteristics and cost parameters were obtained from published studies and hospital databases in Thailand. Costs were expressed in 2014 US Dollars. Outcomes were presented as an incremental cost-effectiveness ratio (ICER). One-way and probabilistic sensitivity analyses were performed to estimate parameter uncertainty.

Results: From a societal perspective, treatment with DPP-4 inhibitors yielded more quality-adjusted life years (QALYs) (0.024) at a higher cost (>66,000 Thai baht (THB) or >1,829.27 USD) per person than SFU, resulting in the ICER of >2.7 million THB/QALY (>74,833.70 USD/QALY). The cost-effectiveness results were mainly driven by differences in HbA1c reduction, hypoglycemic events, and drug acquisition cost of DPP-4 inhibitors. At the ceiling ratio of 160,000 THB/QALY (4,434.59 USD/QALY), the probability that DPP-4 inhibitors are cost-effective compared to SFU was less than 10%.

Conclusions: Compared to SFU, DPP-4 inhibitor monotherapy is not a cost-effective treatment for people with T2DM and CKD in Thailand.

Keywords:

• DPP-4 inhibitor
• Chronic kidney disease
• Diabetes
• Cost-effective
• Thailand

Introduction

Chronic kidney disease (CKD) is common in patients with diabetes¹. A reported prevalence of CKD in type 2 diabetes (T2DM) in primary care diabetes centers in Thailand was 25–27% for CKD stage 3–5 in 2007². A higher prevalence was reported in a tertiary care hospital in Thailand³. According to the Thailand Diabetes Registry (TDR) Project, a multi-center, tertiary hospital-based diabetes registry data of 4,875 people with T2DM, the prevalence of diabetic nephropathy was 42.9% (microalbuminuria 19.7% and overt nephropathy 23.2%), while the prevalence of renal insufficiency and end-stage renal disease (ESRD) were 7.7% and 0.5%, respectively^3,4.

Due to a reduced glomerular filtration rate (GFR) and resultant accumulation of drugs or their metabolites, anti-hyperglycemic treatment options for people with T2DM and CKD are limited⁵. Metformin, the most commonly used anti-hyperglycemia agent, is excreted renally and contraindicated in people with T2DM whose creatinine clearance is less than 30 ml/min⁶. Sulfonylureas (SFU), such as glipizide and gliclazide, can be used; however, these medicines are associated with an increased risk of hypoglycemia and weight gain^7,8. Thiazolidinediones are limited in their use among diabetes people with ESRD because they were associated with increased risks of fluid retention and edema⁹. Although insulin could be considered as an alternative, its doses are needed to be closely monitored to balance between achieving glycemic control and avoiding risk of adverse events, such as hypoglycemia⁷.

A group of anti-hyperglycemia drugs, namely dipeptidyl peptidase-4 (DPP-4) inhibitors, has been recognized as a new avenue for diabetes care in patients with renal impairment because they have been shown to be safe, effective, and well-tolerated. However, the extra benefit gained from the use of DPP-4 inhibitors is needed to justify their additional costs. An economic evaluation, hence, is required to help policy-makers assess the value for money of DPP-4 inhibitor monotherapy compared to SFU monotherapy in people with T2DM and CKD. In Thailand, cost-effectiveness has been recognized as part of the key evidence used for selecting medications to be listed in the National List of Essential Medicine (NLEM), which has been used as a pharmaceutical benefit package for three public health insurance schemes, including Civil Servants Medical Benefits Scheme (CSMBS) for government employees and their dependents, the Social Security Scheme (SSS) for private employees, and the Universal Health Coverage Scheme (UHCS) for the remaining population. This study aimed to assess whether DPP-4 inhibitor monotherapy was cost-effective in people with T2DM and CKD in a Thai context.

Methods

Setting

In Thailand, the health technology assessment (HTA) process used to support the selection of the pharmaceutical reimbursement list involves several steps. The list of HTA topics is initially prioritized by the sub-committee for the development of the NLEM based on the burden of disease and the degree of severity. The sub-committee contracts a Health Economic Working Group (HEWG) to conduct an economic evaluation. The quality of the economic evidence is assured by a comprehensive review of the study report by experts in the field. After the study approval, the findings will be presented to the sub-committee for consideration.

In 2014, DPP-4 inhibitors (saxagliptin, sitagliptin, and vildagliptin) were chosen as one of the listed topics because it is unclear whether these medications should be included in the NLEM for patients with T2DM who had uncontrolled blood glucose and were unable to use insulin. During the study process, the research team undertook two consultation panels with a variety of stakeholders, including policy-makers, health professionals, academics, and private sectors. The aim of the first panel was to reach consensus from participating stakeholders in terms of intervention and comparator, and the scope of the study. At the end of the study, the second panel was held to assess the methodological quality and transparency of input parameters, and the robustness of the study findings. All feedback from stakeholders was taken into consideration in the final analysis. The study had been approved by internal and external reviewers before it was presented to the sub-committee for the development of the NLEM.

Study design

A cost-utility analysis was conducted from the Healthcare System perspective using a Markov model. A validated IMS CORE Diabetes Model (CDM) version 8.5 was used to simulate long-term costs and outcomes associated with each treatment option. A lifetime horizon was used. Both costs and outcomes were discounted at a discount rate of 3% per annum, as suggested by the Thai HTA Guideline¹⁰.

IMS CORE diabetes model

Details of the CDM are described elsewhere^11,12. Briefly, the model consists of 15 sub-models of diabetes complications. Each sub-model uses Markov techniques with Monte Carlo simulation in order to simulate accurately the relationship of multiple complications within individual patients. Four main input databases including cohort, clinical, treatment, and economic databases are required for the model. Clinical parameters used in the CDM were obtained from large, longitudinal studies^13–16. Non-specific mortality rate used in the CDM was adjusted with age-specific mortality rate of Thai people¹⁷. We assumed that 100% of our population were Thai. Other cohort characteristics were derived from the Thai published studies^4,18–24 (Table 1). Cost data were mostly based on Thai hospital databases, hospital costs, or published studies^25–30 (Table 2).

Table 1. Baseline demographics, risk factors, baseline clinical complications of the population.

Variables	Mean (SD)	Data sources
Patient demographics
Mean age (years)	61.8 ± 12.4	TDR 2006^4
Duration of diabetes (years)	12.8 ± 8.2	TDR 2006^4
Proportion of males	40.2%	TDR 2006^4
Baseline risk factors
HbA1c level (%)	8.2 ± 2.5	TDR 2006^4
Systolic Blood Pressure (mmHg)	148.7 ± 25.1	TDR 2006^4
Total Cholesterol (mg/dl)	192.9 ± 57.5	TDR 2006^4
High Density Lipoprotein Cholesterol (mg/dl)	49.1 ± 17.8	TDR 2006^4
Low Density Lipoprotein Cholesterol (mg/dl)	109.3 ± 46.0	TDR 2006^4
Triglycerides (mg/dl)	161.8 ± 136.9	TDR 2006^4
Body Mass Index (kg/m^2)	25.5 ± 4.1	Aekplakorn et al.^18
eGFR (ml/min/1.73 m^2)	40.06 ± 15.71	BCRH hospital database
Proportion of smokers	20.0%	TDR 2006^4
Number of cigarettes smoked/day	10.3	National Statistic Office^19
Alcohol consumption (oz/week)	4.55	Center of Alcohol Studies^20
Racial characteristics
Asian/Pacific Islander	100.00%	Assumption
Baseline cardiovascular disease complications
Myocardial Infarction	4.71%	BCRH hospital database
Angina Pectoris	1.47%	BCRH hospital database
Peripheral Vascular Disease	0.23%	BCRH hospital database
Stroke	2.50%	BCRH hospital database
Congestive Heart Failure	15.17%	BCRH hospital database
Atrial Fibrillation	3.53%	BCRH hospital database
Left Ventricular Hypertrophy	0.44%	BCRH hospital database
Baseline renal complications
Microalbuminuria	18.00%	TDR 2006^4
Gross proteinuria	26.10%	TDR 2006^4
End-stage renal disease	0.10%	Nitiyanant et al.^21
Baseline rRetinopathy complications
Background Retinopathy	25.00%	TDR 2006^4
Proliferative Retinopathy	13.70%	TDR 2006^4
Severe Vision Loss	1.50%	TDR 2006^4
Baseline macular edema
Macular edema	2.50%	Supapluksakul et al.^22
Baseline cataract
Cataract	42.80%	TDR 2006^4
Baseline foot ulcer complications
Uninfected Ulcer	5.90%	TDR 2003^23
Infected Ulcer	1.20%	Nitiyanant et al.^21
Healed Ulcer	6.90%	Nitiyanant et al.^21
History of Amputation	1.50%	TDR 2006^4
Baseline neuropathy
Neuropathy	3.24%	BCRH hospital database
Baseline depression
Depression	28.0%	Thaneerat and Tangwongchai^24

BCRH, Buddhachinaraj Regional Hospital.

Table 2. Cost inputs.

Variables	Mean (THB)	SD	Data sources
Management costs
Costs of aspirin	294	159.39	BCRH hospital database
Costs of statin	1,407	4,266.67	BCRH hospital database
Costs of ACEI	1,019	2,528.07	BCRH hospital database
Costs of antidepressant	1,670	3,300.21	BCRH hospital database
Screening cost for MA (microalbuminuria)	320	—	Maharaj Nakorn Chiang Mai Hospital^25
Screening cost for GRP (gross proteinuria)	60	—	Maharaj Nakorn Chiang Mai Hospital^25
Eye screening	129	—	Pornpinatepong^26
Foot screening program	70	—	Standard cost list^27
Direct costs of CVD complications
Cost of MI first year	106,323	129,552.60	BCRH hospital database
Cost of MI second year onward	26,629	41,451.42	BCRH hospital database
Cost of angina first year	60,235	83,594.51	BCRH hospital database
Cost of angina second year onward	19,578	28,308.46	BCRH hospital database
Cost of CHF first year	58,875	79,235.18	BCRH hospital database
Cost of CHF second year onward	25,452	39,122.61	BCRH hospital database
Cost of stroke first year	71,362	—	BCRH hospital database
Cost of stroke second year onward	23,884	2,123.49	BCRH hospital database
Cost of stroke death within 30 days	38,189	41,778	BCRH hospital database
Cost of PVD first year	156,394	276,600.00	BCRH hospital database
Cost of PVD second year onward	50,374	50,253.25	BCRH hospital database
Direct costs of renal complications
Cost of HD first year	452,120	—	Teerawattananon et al.^28
Cost of HD second year onward	428,141	—	Teerawattananon et al.^28
Cost of PD first year	460,129	—	Teerawattananon et al.^28
Cost of PD second year onward	408,080	—	Teerawattananon et al.^28
Cost of RT first year	928,000	—	King Chulalongkorn Memorial Hospital^29
Cost of RT second year onward	429,240	—	King Chulalongkorn Memorial Hospital^29
Direct costs of acute events
Costs of major hypoglycemia	27,856	70,785.76	BCRH hospital database
Cost of ketoacidosis event	13,284	36,398.48	BCRH hospital database
Cost of lactic acidosis event	64,724	97,511.56	BCRH hospital database
Direct costs of eye diseases
Cost of laser treatment	1,920	—	Pornpinatepong^26
Cost of cataract operation	7,000	—	National Health Security Office^30
Cost of blindness-year of onset	30,902	17,675.91	BCRH hospital database
Cost of blindness-following years	18,766	32,900.26	BCRH hospital database
Direct costs of neuropathy and foot complications
Cost of neuropathy first year	24,410	37,763.1	BCRH hospital database
Cost of neuropathy second year onward	18,797	28,631.95	BCRH hospital database
Cost of amputation prosthesis (event based)	48,365	—	BCRH hospital database
Cost of gangrene treatment (yearly-based)	76,950	95,163.4	BCRH hospital database
Cost of infected ulcer	0	—	Assumption
Cost of uninfected ulcer (yearly-based)	53,076	74,776.36	BCRH hospital database

BCRH, Buddhachinaraj Regional Hospital; THB, Thai Baht; SD, Standard deviation.

Simulation cohort and interventions of interest

The population of interest in this study was Thai people with T2DM and severe CKD. The severe CKD was defined as patients with an estimated glomerular filtration rate (eGFR) < 60 ml/min/1.73 m². Baseline demographics, risk factors, and clinical complications of the population were derived from Thai data sources (Table 1).

The intervention of interest was a monotherapy of DPP-4 inhibitors including saxagliptin, sitagliptin, or vildagliptin. Each DPP-4 inhibitor was compared to glipizide. We used glipizide as a comparator based on the consensus of the first panel consultation. This is because glipizide is the most common SFU used in T2DM with CKD in Thailand. Half of the normal daily dose of DPP-4 inhibitors was used in the analysis, as suggested by a previous study³¹ and the consensus of the first panel consultation. Dose was defined as follows: saxagliptin 2.5 mg/day, sitagliptin 50 mg/day, and vildagliptin 50 mg/day, while dose of glipizide was 10 mg/day.

Efficacy of DPP-4 inhibitors and sulfonylurea

We performed a systematic review and meta-analysis to pool the efficacy of DPP-4 inhibitor monotherapy compared with SFU monotherapy in people with T2DM and CKD (Appendix). The systematic search was performed through MEDLINE, EMBASE, and Clinicaltrial.gov databases from their inception to August 2014. We found three randomized controlled trials (RCTs) that were conducted to determine the efficacy of sitagliptin monotherapy compared to glipizide monotherapy in people with T2DM and CKD^32–34. A meta-analysis using a random-effects model was used to pool the effect size of those three RCTs^32–34. The weighted mean difference (WMD) of HbA1_C reduction between sitagliptin and glipizide was −0.03% (95% CI = −0.28% to 0.22%). Based on the similar efficacy and safety of sitagliptin, saxagliptin, and vildagliptin as treatment for T2DM, either as monotherapy or combination therapy^35–38, the pooled effect size of sitagliptin was assumed for other DPP-4 inhibitors (saxagliptin and vildagliptin). The reduction of HbA1c from SFU monotherapy in T2DM with CKD was also pooled from those three RCTs^32–34 and was equal to −0.72% (95% CI = −0.52% to −0.92%), compared to placebo. Based on the abovementioned WMD and HbA1c reduction of SFU data, HbA1c reduction of DPP-4 inhibitor was estimated to be −0.75% (95% CI = −0.50% to −1.00%), compared to placebo (Table 3).

Table 3. Efficacy and adverse effect of DPP-4 inhibitors and sulfonylurea.

Variables	Mean (95% CI)	Data sources
Efficacy
HbA1c reduction (%)
SFU vs placebo	−0.72 (−0.52 to −0.92)	Pool analysis^32–34
DPP-4 inhibitors vs placebo	−0.75 (−0.50 to −1.00)	Calculation^a
Weight mean difference (DPP-4 inhibitor vs SFU)	−0.03 (−0.28 to 0.22)	Meta-analysis^32–34
Adverse effect
Rate of severe hypoglycemia (times/100 person-year)
SFU	4.06	BCRH hospital database^b
DPP-4 inhibitors	1.36 (0.32 to 5.99)	Calculation^c
Rate of hypoglycemia (times/100 person-year)
SFU	37.94	RECAP-DM study^39 BCRH hospital database^d
DPP-4 inhibitors	13.50 (7.54 to 24.57)	Calculation^e
Risk ratio of severe hypoglycemia (DPP-4 inhibitors vs SFU)	0.34 (0.08 to 1.46)	Meta-analysis
Risk ratio of symptomatic hypoglycemia (DPP-4 inhibitors vs SFU)	0.40 (0.23 to 0.69)	Meta-analysis

^a HbA1c reduction from baseline of DPP-4 inhibitors = −0.72 + (−0.03) = −0.75; Upper 95% CI = −0.72 + (−0.28) = −1.0; Lower 95% CI = −0.72 + (0.22) = −0.5.

^b Risk of severe hypoglycemia of SFU =3.98%, Rate = −ln(1 − 3.98%)*100 = 4.06.

^c Risk of severe hypoglycemia of DPP-4 inhibitors =1.35% (3.98%*0.34), Rate = −ln(1 − 1.35%)*100 = 1.36; Upper 95% CI =5.81% (3.98%*1.46), Rate = −ln(1 − 5.81%)*100 = 5.99; Lower 95% CI =0.32% (3.98%*0.08), Rate = −ln(1 − 0.32%)*100 = 0.32.

^d Risk of hypoglycemia of SFU =31.57%, Rate = −ln(1 − 31.57%)*100 = 37.94.

^e Risk of hypoglycemia of DPP-4 inhibitors =12.63% (31.57%*0.40), Rate = −ln(1 − 12.63%)*100 = 13.50; Upper 95% CI =21.78% (31.57%*0.69), Rate = −ln(1 − 21.78%)*100 = 24.57; Lower 95%CI =7.26% (31.57%*0.23), Rate = −ln(1 − 7.26%)*100 = 7.54.

BCRH, Buddhachinaraj Regional Hospital.

Adverse effects of DPP-4 inhibitors and sulfonylurea

We also pooled the risk ratio (RR) of severe hypoglycemia and symptomatic hypoglycemia from three RCTs^32–34. The overall RR of severe hypoglycemia was 0.34 (95% CI =0.08–1.46), while the overall RR of symptomatic hypoglycemia was 0.40 (95% CI =0.23–0.69). A 1-year rate of severe hypoglycemia and symptomatic hypoglycemia of SFU monotherapy in T2DM with CKD was obtained from a retrospective hospital data analysis and the RECAP-DM study³⁹, indicating 4.06 times/100 person-year and 37.94 times/100 person-year for severe and symptomatic hypoglycemia, respectively. We then applied the RR of severe and symptomatic hypoglycemia to calculate the rate of hypoglycemia of DPP-4 inhibitor monotherapy in T2DM with CKD. The calculation yielded the severe hypoglycemia rate of 1.36 times/100 person-year (95% CI = 0.32–5.99) and symptomatic hypoglycemia rate of 13.50 times/100 person-year (95% CI = 7.54–24.57) (Table 3).

Costs

From the Healthcare System perspective, direct medical and direct non-medical costs were supposed to be included in this study according to the Thai HTA Guideline⁴⁰. However, the consensus from the first consultation panel suggested that direct non-medical costs, such as food, accommodation, and transportation, of both treatment strategies were not different in general practice for T2DM with CKD. Only direct medical costs, therefore, were included in this study. All costs were adjusted for inflation with the Thailand consumer price index (CPI)⁴¹ and presented in year 2014. The costs were converted to USD at a rate of 36.08 THB per USD, as of 30 December 2015⁴².

Direct medical costs included the cost of intervention, concurrent medications, diabetic screening, management cost, and treatment complication. The complications were composed of cardiovascular disease, renal, acute event (hypoglycemia, ketoacidosis, lactic acidosis, and edema), eye disease, neuropathy, and foot complication. Cost data were derived from the published literature and retrospective hospital database analyses^25–30. The details of all cost data are reported in Table 2.

The acquisition cost of DPP-4 inhibitors was calculated based on the unit costs that pharmaceutical companies proposed to the sub-committee for the development for the Development of the NLEM. A package of sitagliptin 100 mg for 28 tablets cost 1,271.16 THB or 35.23 USD. As the recommended dose was 50 mg/day, total cost of sitagliptin was then 8,285.24 THB/year (229.64 USD/year). Vildagliptin 50 mg for 56 tables cost 1,219.80 THB (33.81 USD). The recommended dose was 50 mg/day. Thus, the total cost of vildagliptin was 7,950.48 THB/year (220.36 USD/year). The cost of saxagliptin 5 mg for 28 tablets was 1,035 THB (28.69 USD). The recommended dose was 2.5 mg/day. Thus, total cost of saxagliptin was 6,745.98 THB/year (186.97 USD/year).

The cost of glipizide was derived from a national website of healthcare price from the Ministry of Public Health (www.dmsic.moph.go.th)⁴³. Because there were 14 generics of glipizide available in Thailand to date, we followed the Thai HTA Guideline⁴⁰ and used a median of the median price of those 14 generics, which was 148.68 THB/year (4.12 USD/year) (Table 2).

Sensitivity analyses

To investigate the robustness of the cost-effectiveness results, various types of sensitivity analyses were performed. We conducted a series of one-way sensitivity analysis to examine the parameter uncertainty. These parameters included the change in HbA1c, the rate of hypoglycemia, the cost of DPP-4 inhibitors, and a discount rate. These parameters were varied based on the range of 95% CIs. Annual cost of DPP-4 inhibitors was assumed to be varied by 20% from their mean value. A discount rate was varied from 0–6% based on the recommendation of Thai HTA Guideline¹⁰. A threshold analysis was also conducted to determine the appropriate cost of DPP-4 inhibitors which yielded an ICER within the cost-effectiveness threshold currently used in Thailand (160,000 THB/QALY)⁴⁴. Finally, a probabilistic sensitivity analysis (PSA) was also performed and presented as the cost-effectiveness acceptability curve (CEAC).

Results

Base-case analysis

Table 4 displays the summary findings of the study. Compared with SFU, all three DPP-4 inhibitors yielded equal 0.024 QALYs gained, but had different incremental costs. Saxagliptin incurred the lowest incremental costs, followed by vildagliptin, and sitagliptin (66,481 THB (1,842.60 USD), 80,618 THB (2,234.42 USD), and 84,547 THB (2,343.32 USD), respectively). Therefore, saxagliptin yielded the lowest ICER of 2,747,151 THB/QALY (76,140.55 USD/QALY), followed by vildagliptin (3,331,309 THB/QALY or 92,331.18 USD/QALY), and sitagliptin (3,493,661 THB/QALY or 96,830.96 USD/QALY). Compared with the Thai ceiling threshold, all DPP-4 inhibitors were not cost-effective. Drug acquisition cost was a main cost driver of the incremental cost between the DPP-4 inhibitors and SFU.

Table 4. Results of the base case analysis.

Treatment	Total cost THB (USD)	Quality-adjusted life year	Incremental costs THB (USD)	Incremental effectiveness (QALYs gained)	Incremental cost-effectiveness ratio, THB/QALY (USD/QALY)
Saxagliptin vs Sulfonylurea
Saxagliptin	624,797 (17,316.99)	7.552	66,481 (1,842.60)	0.024	2,747,151 (76,140.55)
SFU	558,315 (15,474.36)	7.528
Sitagliptin vs Sulfonylurea
Sitagliptin	642,862 (17,817.68)	7.552	84,547 (2,343.32)	0.024	3,493,661 (96,830.96)
SFU	558,315 (15,474.36)	7.528
Vildagliptin vs Sulfonylurea
Vildagliptin	638,933 (17,708.79)	7.552	80,618 (2,234.42)	0.024	3,331,309 (92,331.18)
SFU	558,315 (15,474.36)	7.528

Sensitivity analyses

Of those comparisons between DPP-4 inhibitors and SFU, saxagliptin showed the lowest ICER in the base case analysis; hence, the results of one-way sensitivity analysis were based entirely on saxagliptin compared with SFU. Figure 1 shows that the ICER was most sensitive to the effect of DPP-4 inhibitors on HbA1c reduction, discount rate, severe hypoglycemia rate, symptomatic hypoglycemia rate, and annual cost of saxagliptin. It was found that ICERs were still above the ceiling ratio of 160,000 THB/QALY with 80% reduction in the annual cost of DPP-4 inhibitors from the base case analysis (Figure 2). Figure 3 displays an ICER scatterplot of saxagliptin compared with SFU on a cost-effectiveness plan. Most ICERs fell in the upper left and right quadrants.

Figure 1. Tornado diagram of saxagliptin compared with sulfonylurea in type 2 diabetes with chronic kidney disease.

Figure 2. Threshold analysis of DPP-4 inhibitors vs sulfonylurea in type 2 diabetes with chronic kidney disease.

Figure 3. Incremental cost-effectiveness ratio scatterplot of saxagliptin vs sulfonylurea in type 2 diabetes with chronic kidney disease.

At the ceiling ratio of 160,000 THB/QALY, DPP-4 inhibitors had less than 10% of being cost-effective compared with SFU (Figure 4). Saxagliptin (5.9%) had the highest probability of being a cost-effective alternative, followed by vildagliptin (2.7%), and sitagliptin (1.9%), respectively.

Figure 4. Cost-effectiveness acceptability curve of DPP-4 inhibitors vs sulfonylurea in type 2 diabetes with chronic kidney disease.

Discussion

T2DM treatments are costly, especially for individuals with comorbidities, such as renal impairment. This study shows the long-term costs and effectiveness of DPP-4 inhibitor monotherapy compared to SFU monotherapy in people with T2DM and CKD in the Thai context. Our analysis was based on the data obtained from a representative sample of target population of DPP-4 inhibitors in Thailand. Compared to SFU, we found that DPP-4 inhibitors was associated with improved health outcomes, i.e. greater value of 0.024 QALYs and fewer diabetes-related complications relative to SFU (RR of severe hypoglycemia = 0.34, RR of symptomatic hypoglycaemia =0.40). However, therapeutic benefits and associated cost-savings do not outweigh the higher acquisition cost of DPP-4 inhibitors. Compared to SFU, DPP-4 inhibitors were, therefore, not cost-effective at a currently used threshold in Thailand, suggesting that the use of DPP-4 inhibitors for people with T2DM and CKD in Thailand is unlikely to represent an efficient use of finite healthcare budget.

To our knowledge, our study is the first cost-effectiveness study of DPP-4 inhibitors that compared the DPP-4 inhibitor monotherapy with SFU monotherapy in people with T2DM and CKD based on a recently published systematic review of DPP-4 inhibitor for T2DM⁴⁵. Previous studies focused on the use DPP-4 inhibitors, including saxagliptin and sitagliptin, plus metformin with sulfonylurea plus metformin as second-line therapy^46–50. The conclusion of these studies all suggested that the use of DPP-4 inhibitors plus metformin as a second-line add-on therapy was cost-effective when compared with a SFU plus metformin in T2DM.

The reason how this cost-effectiveness question was raised for the CKD population deserves discussion. During a stakeholder meeting for a topic refinement, various target populations were discussed. CKD population was chosen because limited treatment options are available. DPP-4 inhibitors, therefore, have a potential role for this population. In addition, the size of population with CKD is small, making the budget implication potentially affordable from a policy decision perspective.

There are some limitations regarding the availability of data used in the study. First, the study relies on limited evidence for the efficacy parameters from our unpublished systematic review and meta-analysis, which demonstrated no statistically significant difference in terms of HbA1c reduction. Second, although this study had advantages in using hospital data of Thai people with T2DM and CKD to estimate the baseline clinical data, and severe hypoglycemia, the information on the symptomatic hypoglycemia was derived from the RECAP-DM study³⁹, which included Thai T2DM with or without CKD who treated with SFU monotherapy. We were unable to estimate the rate of symptomatic hypoglycemia from the actual study population because people with T2DM who experienced symptomatic hypoglycemia were less likely to visit or be admitted in a hospital. Third, there were ∼52-times greater acquisition costs of DPP-4 inhibitors compared to SFU used in the study. SFU has many generics available in Thailand, while DPP-4 inhibitors are branded drugs. It is challenging for a new intervention to show value for money when comparing to an alternative that has much lower acquisition cost. However, we believe that our findings would lead to the pharmaceutical companies’ decision to reduce drug price. Moreover, some cost data, such as infected ulcer, were not available, these costs were assumed to be zero. However, we found negligible impact of these cost data on the cost-effective findings. For example, when cost of infected ulcer was assumed to be similar to uninfected ulcer, an ICER of saxagliptin compared to SFU would decrease slightly from 2,747,151 THB/QALY (76,140.55 USD/QALY) to 2,746,400 THB/QALY (76,119.73 USD/QALY). Third, utility values, transition probabilities between different health states within the CDM were mostly based on studies conducted in other countries. Without existing well-established studies in people with T2DM and CKD, we believe that we performed an extensive search in order to seek for the most credible evidence and used very conservative assumptions where information was limited. Although these parameters were not addressed in sensitivity analyses, we believe that our findings would not change due to several reasons. First, the results from the base-case analysis showed an extremely high ICER value (> 2.7 million THB) compared with the threshold value of 160,000 THB/QALY for DPP-4 inhibitor monotherapy compared with SFU monotherapy. Second, the efficacy of DPP-4 inhibitor monotherapy was not significantly different from that of SFU (WMD = −0.03, CI = −0.28 to 0.22); therefore, estimated QALY of both alternatives were not substantially different. As shown in our one-way sensitivity analyses, the main drivers of our study finding were the efficacy of DPP-4 and cost, especially drug cost.

Another major limitation of our study is that we have not taken into account the cardiovascular risk associated with the use of DPP-4 inhibitors and SFU. The concern of increased cardiovascular risk among users of these two groups of products remained unresolved, despite a large number of research attempting to investigate this issue^51–53. Based on this limitation, the findings of our cost-effectiveness study should be interpreted with caution.

Conclusion

Compared to SFU monotherapy, DPP-4 inhibitor monotherapy is not a cost-effective strategy for people with T2DM and CKD in the era of limited healthcare resources in Thailand. Given that the increasing prevalence of people with T2DM and CKD and high acquisition costs for DPP-4 inhibitors, the budgetary impact analysis for this population should be considered.

Transparency

Declaration of funding

This study was supported by grants from the Subcommittees of the National List of Essential Medicine, Ministry of Public Health, Thailand.

Declaration of financial/other relationships

The authors and JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

The authors gratefully acknowledge the support of the National Drug Selection Working Group in Endocrinology for their time and suggestions, and the IMS Health team for supporting the IMS CORE Diabetes Model.

Appendix

Systematic review and meta-analysis of dipeptidyl peptidase-4 inhibitor monotherapy and sulfonyl urea monotherapy for people with type 2 diabetes and chronic kidney disease in Thailand

Method

Data sources and searching strategy

The following databases were systematically searched: MEDLINE, EMBASE, and Clinicaltrial.gov. Databases were searched from their inception to August 12, 2014. For the search strategy, we used

(“new user” OR “drug-naive” OR “Monotherapy”) AND (Sulfonylurea OR gliclazide OR glyburide OR glibenclamide OR glipizide OR glimepiride OR glycopyramide OR glisoxipide OR gliquidone OR glibornuride OR carbutamide OR acetohexamide OR chlorpropamide OR tolbutamide) AND (“Dipeptidyl-Peptidase IV Inhibitors” OR sitagliptin OR saxagliptin OR vildagliptin OR linagliptin OR anagliptin OR teneligliptin OR alogliptin)

References of initially identified articles will be further examined to identify additional studies that met the selection criteria.

Study selection

We included studies that: (i) compared monotherapy of sulfonylurea vs DPP-4 Inhibitors, (ii) reported the number (or percentage with total number of patients) of outcomes, and (iii) studied humans. There was no language restriction. Abstracts/titles and full-text articles were reviewed by two independent reviewers. We excluded studies that were not original articles such as comments, letters, reviews, meta-analyses, guidelines, case reports, surveys, or editorials. Studies from the same population (duplicate studies), and studies not reporting effect-estimates or with insufficient information to compute effect estimates were also excluded.

Results

Flow diagram of study selection.