Economic outcomes with prasugrel versus clopidogrel in acute coronary syndrome patients: observations from prasugrel users and matched clopidogrel users

Figures & data

Table 1. Demographics and clinical characteristics of clopidogrel/prasugrel pre-matched and matched-cohorts.

Variable	Pre-matched cohorts			Matched cohorts*
	Clopidogrel	Prasugrel	Std Diff†,‡	Clopidogrel	Prasugrel	Std Diff†,‡
Number of patients	65,392	10,963		10,963	10,963
Demographics§
Age, mean (SD)	63.8 (12.7)	57.4 (9.2)	57.6%	57.4 (10.6)	57.4 (9.2)	0.0%
Female, n (%)	20,836 (31.9)	2403 (21.9)	22.4%	2360 (21.5)	2403 (21.9)	1.0%
Region§, n (%)
South	24,233 (37.1)	4415 (40.3)	6.6%	4440 (40.5)	4415 (40.3)	0.5%
Northeast	9284 (14.2)	1461 (13.3)	2.5%	1457 (13.3)	1461 (13.3)	0.1%
West	10,066 (15.4)	1717 (15.7)	0.7%	1680 (15.3)	1717 (15.7)	0.9%
North Central	21,297 (32.6)	3216 (29.3)	7.0%	3228 (29.4)	3216 (29.3)	0.2%
Unknown	512 (0.8)	154 (1.4)	6.0%	158 (1.4)	154 (1.4)	0.3%
Type of insurance§, n (%)
Comprehensive	14,177 (21.7)	1262 (11.5)	27.3%	1209 (11.0)	1262 (11.5)	1.5%
HMO	9353 (14.3)	1455 (13.3)	3.0%	1437 (13.1)	1455 (13.3)	0.5%
POS	4145 (6.3)	833 (7.6)	4.9%	850 (7.8)	833 (7.6)	0.6%
PPO	33,015 (50.5)	6240 (56.9)	12.9%	6278 (57.3)	6240 (56.9)	0.7%
Other	3366 (5.1)	913 (8.3)	12.7%	915 (8.3)	913 (8.3)	0.1%
Unknown	1336 (2.0)	260 (2.4)	2.2%	274 (2.5)	260 (2.4)	0.8%
Year of index date§, n (%)
2009	13,653 (20.9)	371 (3.4)	53.6%	366 (3.3)	371 (3.4)	0.3%
2010	23,460 (35.9)	3420 (31.2)	9.9%	3525 (32.2)	3420 (31.2)	2.1%
2011	19,993 (30.6)	4675 (42.6)	25.1%	4734 (43.2)	4675 (42.6)	1.1%
2012	8286 (12.7)	2497 (22.8)	26.5%	2338 (21.3)	2497 (22.8)	3.5%
Charlson Comorbidity Index¶, mean (SD)	2.07 (1.69)	1.65 (1.29)	28.2%	1.64 (1.32)	1.65 (1.29)	0.6%
0	6361 (9.7%)	1118 (10.2%)	1.6%	1200 (10.9%)	1118 (10.2%)	2.4%
1	24,215 (37.0%)	5217 (47.6%)	21.4%	5200 (47.4%)	5217 (47.6%)	0.3%
2	15,859 (24.3%)	2662 (24.3%)	0.1%	2593 (23.7%)	2662 (24.3%)	1.5%
3	8685 (13.3%)	1154 (10.5%)	8.5%	1149 (10.5%)	1154 (10.5%)	0.1%
4	4601 (7.0%)	443 (4.0%)	13.1%	457 (4.2%)	443 (4.0%)	0.6%
5 or more	5671 (8.7%)	369 (3.4%)	22.3%	364 (3.3%)	369 (3.4%)	0.3%
Other comorbidities¶, n (%)
Acute myocardial infarction	46,559 (71.2)	8441 (77.0)	13.2%	8363 (76.3)	8441 (77.0)	1.7%
Hypertension	39,929 (61.1)	6090 (55.6)	11.2%	6113 (55.8)	6090 (55.6)	0.4%
Dyslipidemia	27,664 (42.3)	5115 (46.7)	8.8%	5116 (46.7)	5115 (46.7)	0.0%
Diabetes	19,430 (29.7)	2917 (26.6)	6.9%	2866 (26.1)	2917 (26.6)	1.1%
Congestive heart failure	11,042 (16.9)	1157 (10.6)	18.4%	1154 (10.5)	1157 (10.6)	0.1%
Substance abuse (drug or alcohol)	9228 (14.1)	1934 (17.6)	9.7%	1978 (18.0)	1934 (17.6)	1.0%
Ischemic stroke or TIA	7594 (11.6)	465 (4.2)	27.3%	458 (4.2)	465 (4.2)	0.3%
Anemia	7455 (11.4)	584 (5.3)	21.9%	556 (5.1)	584 (5.3)	1.2%
Ventricular arrhythmias	2805 (4.3)	474 (4.3)	0.2%	465 (4.2)	474 (4.3)	0.4%
Cirrhosis/hepatic disease	1784 (2.7)	255 (2.3)	2.6%	271 (2.5)	255 (2.3)	1.0%
End stage renal disease	882 (1.3)	52 (0.5)	9.2%	54 (0.5)	52 (0.5)	0.3%
Hemophilia	473 (0.7)	59 (0.5)	2.3%	75 (0.7)	59 (0.5)	1.9%
Prior hemorrhagic stroke	169 (0.3)	5 (0.0)	5.5%	3 (0.0)	5 (0.0)	1.0%
History of HIV infection	147 (0.2)	43 (0.4)	3.0%	50 (0.5)	43 (0.4)	1.0%
Maternity patient	69 (0.1)	15 (0.1)	0.9%	18 (0.2)	15 (0.1)	0.7%
Procedure¶, n (%)
PCI	47,036 (71.9)	10,475 (95.5)	64.0%	10,475 (95.5)	10,475 (95.5)	0.0%
Baseline medication¶, n (%)
Anti-hypertensive	30,009 (45.9)	4599 (42.0)	7.9%	4526 (41.3)	4599 (42.0)	1.4%
Anti-hyperlipidemic	31,369 (48.0)	5070 (46.2)	3.5%	5026 (45.8)	5070 (46.2)	0.8%
Anti-coagulant	3243 (5.0)	225 (2.1)	15.8%	196 (1.8)	225 (2.1)	1.9%

HIV, human immunodeficiency virus; HMO, health maintenance organization; PCI, percutaneous coronary intervention; POS, point of service; PPO, preferred provider organization; Std Diff, standardized difference; SD, standard deviation; TIA, transient ischemic attack.

*Clopidogrel-treated patients were matched exactly 1:1 to prasugrel patients on the following variables: propensity score percentiles (5th, 10th, 15th, 90th, 95th, and 100th), and PCI.

†For continuous variables, the standardized difference is calculated by dividing the absolute difference in means of the clopidogrel and the prasugrel cohorts by the pooled standard deviation of both groups. The pooled standard deviation is the square root of the average of the squared standard deviations.

‡For binary outcomes, the standardized difference is calculated using the equation below, where P is the respective proportion of participants in each group: (Pclopidogrel − Pprasugrel)/√[p(1 − p)], where p = (Pclopidogrel + Pprasugrel)/2.

§Evaluated at the index date.

¶Evaluated during the 6-month baseline period. This table has been reproduced with permission from Olson et al.26. © 2014 John Wiley & Sons Ltd.

Table 2. Strength of evidence using Bayes factors21.

Strength of evidence	Evidence against H0	Evidence for H0
Not worth more than a bare mention	1–3	0.33–1
Positive	3–20	0.05–0.33
Strong	20–150	0.0067–0.05
Very strong	>150	<0.0067

H₀, hypothesis in the denominator. This table has been reproduced with permission from Olson et al.26. © 2014 John Wiley & Sons Ltd.

Figure 1. Illustration of the five key posterior probabilities. r, ratio of means; d, non-inferiority/equivalence margin. The posterior probabilities are the areas under the posterior density curve for the indicated intervals. This figure has been reproduced with permission from Olson et al.26. © 2014 John Wiley & Sons Ltd.

Table 3. Mean ACS-related healthcare cost and all-cause healthcare costs per patient for up to 30 and 365 days of follow-up for clopidogrel/prasugrel matched-cohort.

Mean healthcare cost incurred over 30/365 days (2011 $USD)	Time period, days	Clopidogrel vs prasugrel
		Clopidogrel (n = 10,963)	Prasugrel (n = 10,963)	Cost difference* (95% CI)
ACS-related healthcare
Hospitalizations	30	$1500	$1587	−$87 (−$399; $196)
	365	$6137	$5930	$207 (−$615; $946)
ER visits	30	$83	$85	−$1 (−$19; $16)
	365	$305	$290	$14 (−$35; $65)
Outpatient visits	30	$1337	$1538	−$201 (−$337; −$67)†
	365	$5839	$6158	−$319 (−$805; $183)
All-cause healthcare
Hospitalizations	30	$1542	$1620	−$78 (−$393; $206)
	365	$6635	$6317	$318 (−$527; $1063)
ER visits	30	$137	$137	$1 (−$21; $21)
	365	$571	$587	−$16 (−$131; $74)
Outpatient visits	30	$1740	$1893	−$153 (−$305; −$8)†
	365	$9857	$9748	$109 (−$554; $754)

CI, confidential interval; ER, emergency room.

*Difference in cost for clopidogrel-treated patients and prasugrel-treated patients; negative numbers indicate lower costs with clopidogrel.

†Statistically significant.

Mean healthcare costs were calculated with the Lin’s cost history method, assuming that all patients were censored at the end of follow-up and had no event that terminated cost accumulation (e.g., death). The Kaplan–Meier estimates thus remained equal to 1 in Lin’s cost history method calculations.

The 95% CI was based on non-parametric bootstrap samples using 1000 replications.

Figure 2. Posterior densities for ratio of means of ACS-related hospitalization costs at 30 days and 365 days. *A ratio of <1 indicates that the mean cost of clopidogrel is lower than the mean cost of prasugrel.

Table 4. Bayesian analysis: posterior distribution of ACS-related mean healthcare cost and all-cause mean healthcare costs for matched-cohorts.

Type of cost	Time period, days^a,b	Clopidogrel		Prasugrel		Ratio of mean cost
		Mean (SD) ($)	CI*	Mean (SD) ($)	CI*	Mean (SD)	CI*
ACS related mean healthcare
Hospitalizations	30	1489 (36)	1422–1562	1603 (40)	1527–1679	0.930 (0.032)	0.873–0.999
	365	4485 (216)	4089–4929	4754 (252)	4275–5269	0.946 (0.068)	0.821–1.086
ER visits	30	84 (5)	75–94	87 (5)	77–97	0.978 (0.083)	0.819–1.153
	365	294 (21)	254–336	294 (19)	258–334	1.001 (0.095)	0.828–1.205
Outpatient visits	30	1362 (23)	1317–1407	1557 (27)	1504–1614	0.875 (0.021)	0.833–0.917
	365	5622 (140)	5354–5895	6382 (177)	6047–6765	0.882 (0.033)	0.814–0.947
All-cause mean healthcare
Hospitalizations	30	1531 (39)	1457–1609	1638 (41)	1562–1718	0.935 (0.033)	0.873–1.003
	365	4870 (228)	4442–5348	5023 (269)	4535–5583	0.972 (0.071)	0.829–1.115
ER visits	30	139 (6)	127–151	139 (6)	128–151	1.003 (0.06)	0.883–1.121
	365	552 (26)	503–605	534 (28)	483–587	1.038 (0.075)	0.899–1.199
Outpatient visits	30	1770 (25)	1721–1818	1934 (27)	1880–1989	0.915 (0.018)	0.879–0.951
	365	9492 (196)	9120–9880	9989 (218)	9574–10,425	0.951 (0.028)	0.895–1.006

ACS, acute coronary syndrome; CI, credible interval; ER, emergency room; SD, standard deviation.

*CI, 2.5% lower bound; 97.5% upper bound.

Healthcare costs were modeled using a zero-inflated Gamma distribution.

^aA total of 10,562 clopidogrel patients and 10,584 prasugrel patients were included in the 30-day analysis.

^bA total of 2743 clopidogrel patients and 2366 prasugrel patients were included in the 365-day analysis.

Table 5. Posterior superiority probabilities and superiority Bayes factors for ratio of means of healthcare cost at 30 days (clopidogrel mean/prasugrel mean; clopidogrel, n = 10,562; prasugrel, n = 10,584).

Type of cost	Hypothesis	Posterior probabilities	Bayes factor (C superior/P superior)
All-cause hospitalization	Clopidogrel superior	0.969	31.258
	Prasugrel superior	0.031
All-cause ER	Clopidogrel superior	0.484	0.938
	Prasugrel superior	0.516
All-cause outpatient	Clopidogrel superior	1.000	∞
	Prasugrel superior	0
ACS-related hospitalization	Clopidogrel superior	0.976	40.667
	Prasugrel superior	0.024
ACS-related ER	Clopidogrel superior	0.613	1.584
	Prasugrel superior	0.387
ACS-related outpatient	Clopidogrel superior	1.000	∞
	Prasugrel superior	0

ACS, acute coronary syndrome; ER, emergency room.

Table 6. Posterior superiority probabilities and superiority Bayes factors for ratio of means of healthcare cost at 365 days (clopidogrel mean/prasugrel mean; clopidogrel, n = 2743; prasugrel, n = 2366).

Type of cost	Hypothesis	Posterior probabilities	Bayes factor (C superior/P superior)
All-cause hospitalization	Clopidogrel superior	0.662	1.959
	Prasugrel superior	0.338
All-cause ER	Clopidogrel superior	0.324	0.479
	Prasugrel superior	0.676
All-cause outpatient	Clopidogrel superior	0.964	26.778
	Prasugrel superior	0.036
ACS-related hospitalization	Clopidogrel superior	0.800	4.000
	Prasugrel superior	0.200
ACS-related ER	Clopidogrel superior	0.536	1.158
	Prasugrel superior	0.463
ACS-related outpatient	Clopidogrel superior	1.000	∞
	Prasugrel superior	0

ACS, acute coronary syndrome; ER, emergency room.

Table 7. Posterior equivalence and non-inferiority probabilities of healthcare cost at 30 days based on posterior distribution of ratio of means (clopidogrel mean/prasugrel mean; clopidogrel, n = 10,562; prasugrel, n = 10,584).

Type of cost	Clopidogrel vs prasugrel	Margins
		0.05	0.10	0.15	0.20
All-cause hospitalization	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.320	0.860	0.997	1.000
	Clopidogrel non-inferior	1.000	1.000	1.000	1.000
	Prasugrel non-inferior	0.320	0.860	0.997	0.969
All-cause ER	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.599	0.900	0.985	0.999
	Clopidogrel non-inferior	0.789	0.944	0.991	0.999
	Prasugrel non-inferior	0. 810	0.956	0.994	1.000
All-cause outpatient	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.030	0.802	1.000	1.000
	Clopidogrel non-inferior	1.000	1.000	1.000	1.000
	Prasugrel non-inferior	0.030	0.802	1.000	1.000
ACS-related hospitalization	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.264	0.819	0.998	1.000
	Clopidogrel non-inferior	1.000	1.000	1.000	1.000
	Prasugrel non-inferior	0.264	0.819	0.998	1.000
ACS-related ER	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.442	0.762	0.913	0.977
	Clopidogrel non-inferior	0.816	0.935	0.973	0.993
	Prasugrel non-inferior	0.626	0.827	0.940	0.984
ACS-related outpatient	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.000	0.117	0.877	1.000
	Clopidogrel non-inferior	1.000	1.000	1.000	1.000
	Prasugrel non-inferior	0.000	0.117	0.877	1.000

ACS, acute coronary syndrome; ER, emergency room.

Table 8. Posterior equivalence and non-inferiority probabilities for ratio of means of healthcare cost at 365 days (clopidogrel mean/prasugrel mean; clopidogrel, n = 2743; prasugrel, n = 2366).

Type of cost	Clopidogrel vs prasugrel	Margins
		0.05	0.10	0.15	0.20
All-cause hospitalization	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.500	0.807	0.950	0.992
	Clopidogrel non-inferior	0.871	0.959	0.991	0.999
	Prasugrel non-inferior	0.629	0.848	0.960	0.993
All-cause ER	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.477	0.777	0.917	0.976
	Clopidogrel non-inferior	0.591	0.803	0.919	0.976
	Prasugrel non-inferior	0.886	0.974	0.997	1.000
All-cause outpatient	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.525	0.956	1.000	1.000
	Clopidogrel non-inferior	1.000	1.000	1.000	1.000
	Prasugrel non-inferior	0.525	0.956	1.000	1.000
ACS-related hospitalization	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.401	0.736	0.926	0.988
	Clopidogrel non-inferior	0.937	0.983	0.996	0.999
	Prasugrel non-inferior	0.464	0.754	0.931	0.988
ACS-related ER	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.400	0.707	0.891	0.959
	Clopidogrel non-inferior	0.713	0.844	0.927	0.969
	Prasugrel non-inferior	0.686	0.863	0.964	0.990
ACS-related outpatient	Posterior probabilities
	Clopidogrel is equivalent to prasugrel	0.021	0.294	0.823	0.990
	Clopidogrel non-inferior	1.000	1.000	1.000	1.000
	Prasugrel non-inferior	0.021	0.294	0.823	0.990

ACS, acute coronary syndrome; ER, emergency room.

Olson WH, Ma Y-W, Laliberté F, et al. Prasugrel vs. clopidogrel in acute coronary syndrome patients treated with prasugrel. J Clin Pharm Ther 2014;39:663-72

 PubMed Web of Science ®Google Scholar

Amsterdam EA, Wenger NK, Brindi RG, et al. 2014 AHA/ACC guideline for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 2014;130:e344-426

 PubMed Web of Science ®Google Scholar