|Perioperative anticoagulation for patients with mechanical heart valves: a model comparing unfractionated and low-molecular-weight heparin
|Garcia D A, Libby E N, Rich J S
This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.
The use of intravenous (IV) unfractionated heparin (UFH) and low molecular weight heparins (LMWH) for short-term perioperative anticoagulation in patients with mechanical heart valves undergoing invasive procedures.
The authors built a model to analyse the costs and consequences for a population of 55-year-old men and women with a St. Jude mitral valve and a history of atrial fibrillation, who were receiving perioperative antithrombotic agents when referred for invasive procedures.
The setting was secondary care. The economic study was carried out in the USA.
Dates to which data relate
The effectiveness evidence came from studies published between 1975 and 2002. The resource use data were drawn from a study published in 1990. The costs were obtained at the time of analysis in the authors' hospital setting. The dates to which the prices related were not reported.
Source of effectiveness data
The effectiveness data were derived from a review or synthesis of studies, and estimates of effectiveness based on opinion.
A Markov model, developed using DATA 3.5 software, was used to estimate the quality-adjusted life-years (QALYs) and health care costs over the patient's lifetime. Further details of the model were not reported.
Outcomes assessed in the review
Model input parameters drawn from literature comprised the following:
the annual risk of a major thromboembolic event,
the annual risk of a stroke,
the annual risk of mitral valve thrombosis,
the annual risk of a peripheral embolic event,
the probability of dying from acute stroke,
the probability of a permanent neurologic deficit,
the probability of dying from valve thrombosis,
the probability of dying from an acute peripheral embolic event, and
the probability of permanent tissue loss (amputation or bowel resection).
Study designs and other criteria for inclusion in the review
Sources searched to identify primary studies
Criteria used to ensure the validity of primary studies
Methods used to judge relevance and validity, and for extracting data
Number of primary studies included
Ten primary studies or references provided effectiveness evidence.
Methods of combining primary studies
Most of the primary studies were not combined.
Investigation of differences between primary studies
The authors did not investigate any differences between the primary studies. They did not explain differences between the individual studies, nor did they investigate how these differences affected the estimate of effectiveness.
Results of the review
The annual risks per 100 patient-years of thromboembolic events associated with UFH were 2.3 (range: 1.8 - 6.9) for major thromboembolic events, 1.3 (range: 1.0 - 3.9) for strokes, and 0.5 (range: 0.4 - 1.5) for mitral valve thrombosis and for peripheral embolic events.
Several probabilities were associated with both UFH and LMWH:
the probability of dying from acute stroke was 15% (range: 10 - 40);
the probability of a permanent neurologic deficit was 40% (range: 20 - 60);
the probability of dying from valve thrombosis was 50% (range: 25 - 75);
the probability of dying from an acute peripheral embolic event was 10% (range: 5 - 15); and
the probability of permanent tissue loss was 10% (range: 5 - 15).
Methods used to derive estimates of effectiveness
The authors made assumptions to derive estimates of effectiveness.
Estimates of effectiveness and key assumptions
The daily risk of thromboembolic events was assumed to be 50% higher in patients receiving LMWH than in patients receiving UFH. The assumption was acknowledged as arbitrary and was made in order to favour the UFH strategy. The authors assumed that there was no difference between the two strategies in the risk of serious bleeding, based on a comparative study in postoperative patients and randomised controlled trials in other indications. They also assumed that there was no temporary increase in the risk of thromboembolism related to the abrupt discontinuation of warfarin or the postoperative state, under the rationale that definitive evidence to the contrary was lacking. Finally, the authors assumed that patients surviving valve thrombosis suffered no permanent sequelae and returned to the "well on chronic anticoagulation" health state.
Measure of benefits used in the economic analysis
The measure of benefits used was the QALYs. It was assumed that patients experiencing no adverse events and receiving either warfarin, UFH or LMWH had a utility of 1. Other health states and associated utilities were as follows:
stroke survivor with major neurological deficit, 0.2 (range: 0.2 - 0.8);
stroke survivor with minor neurological deficit, 0.6 (range: 0.6 - 0.9);
peripheral embolic event survivor with amputation or bowel resection, 0.5 (range: 0.2 - 0.8);
any complication during the acute phase, 0.2 (range: 0.2 - 0.5).
With the exception of the last, which was an assumption, all of these were drawn from the literature. The authors also assumed no effect on quality of life of hospitalisation for IV UFH.
An annual discount rate of 3% was applied to the costs and benefits. The authors considered the costs of thromboprophylaxis strategies, as well as acute and long-term costs associated with thromboembolic complications. Daily medication costs (including hospitalisation when necessary) and costs of valve-related complication events were applied in the model. However, the quantities were neither analysed nor reported. The acute and long-term costs of stroke were drawn from a 1990 publication, while costs of medications and emergency valve replacement were estimated from the perspective of the authors' hospital setting (date not reported). Thus, charges were used to approximate the costs. The price year was not reported.
Statistical analysis of costs
The costs were treated deterministically.
The indirect costs were not included.
A one-way sensitivity analysis was carried out. The parameters varied were the relative effectiveness of LMWH, the absolute risk of thromboembolism in the base-case (examining the effect of a possible transient hypercoagulable state related to abrupt warfarin discontinuation), and the cost parameters for hospitalisation. The method used to select the ranges was not specified.
Estimated benefits used in the economic analysis
The total number of QALYs was 13.4422 for LMWH and 13.4427 for IV UFH.
The estimated incremental benefits of UFH over LMWH were 0.0005 QALYs gained per patient.
The total costs associated with UFH were $55,600 per patient ($54,300 if patients with mechanical heart valves were to be hospitalised for other reasons after an invasive procedure), while the total costs associated with LMWH were $53,900 per patient. This translated to a total incremental cost with UFH of $1,600 per patient ($400 if additional hospitalisations).
Synthesis of costs and benefits
The estimated benefits and costs were combined into an incremental cost-effectiveness ratio, the cost per QALY gained. The cost-effectiveness of using UFH compared with LMWH was $3,200,000 per QALY gained. If patients with mechanical heart valves were to be hospitalised, then the cost-effectiveness using UFH compared with LMWH was $840,000 per QALY gained.
Over the range of values examined in the sensitivity analysis, the cost-effectiveness ratios for IV UFH did not fall below $1,000,000 per QALY gained until the risk of any thromboembolic event was more than three times the base-case estimate. If, simultaneously, pessimistic assumptions about hypercoagulability (increase in risk by a factor of 5) and inferiority of LMWH (increase in risk by a factor of 3) were applied in the base-case, the cost-effectiveness ratios for UFH were still more than $100,000 per QALY gained.
The authors reported that, under the assumption that UFH was more effective in preventing valve-related complications, 4,167 patients would have to be hospitalised to receive UFH in order to prevent one major thromboembolic event.
Hospitalising even high-risk patients for intravenous (IV) unfractionated heparin (UFH) is "prohibitively expensive". The cost-effectiveness ratios for IV UFH substantially exceed the values of other widely adopted preventive measures.
CRD COMMENTARY - Selection of comparators
The comparator (LMWH) was justified on the grounds of its increasing use as an alternative to UFH in the hospital setting in the USA. You should decide whether it is a widely used health technology in your own setting.
Validity of estimate of measure of effectiveness
The authors did not state that a systematic review of the literature had been undertaken. The authors used data from the available studies selectively. Uncertainty remains around the literature search and possible omission of relevant studies, as no details of any search strategy were reported in the paper. The impact of differences between the primary studies was not considered when estimating effectiveness. It is important to note that a key input to the model, the effectiveness of LMWH in preventing thromboembolic events, was not drawn from published evidence but was based on an authors' assumption. The authors justified their other assumptions by reference to the medical literature. The estimates were investigated in a sensitivity analysis, but it was unclear whether the literature provided the sources for the ranges used.
Validity of estimate of measure of benefit
The estimate of benefits (QALYs) was modelled. The instrument used to derive a measure of health benefit, a Markov model, was appropriate. The utility values were drawn from the literature, but it was not reported whether the populations in which they were measured were equivalent to the population entering the model. The utility scores were not varied in the sensitivity analysis despite being a key driver of benefits.
Validity of estimate of costs
All the categories of cost relevant to the third-party payer were included in the analysis. The costs and the quantities were not reported separately, which may limit extrapolation exercises to other settings. Resource use and prices were taken from published sources and the authors' own setting. A sensitivity analysis of the costs was conducted. The ranges used were authors' estimates but they would appear to be appropriate. The date to which the costs related was not reported, which will hinder any future inflation exercises. Discounting was appropriately applied as the study horizon was longer than 2 years. Hospital charges were used to proxy prices for hospital care; such charges do not reflect true opportunity costs (due to profit margin) and (in the absence of a cost-to-charge ratio) may limit the generalisability of the results beyond the authors' clinical setting.
The authors did not make appropriate comparisons of their findings with those from other studies. The issue of generalisability to other studies was not addressed and was made more difficult by the lack of transparency in the prices and quantities of resource use. The authors did not present their results selectively. The authors' conclusions reflected the scope of the analysis in judging the suitability of LMWH as an alternative strategy in the USA for patients needing perioperative anticoagulation. It is interesting that, despite a number of assumptions designed to favour UFH, the therapy showed unacceptably high cost-per-QALY ratios in comparison with the newer less proven therapy, LMWH.
The authors reported two limitations to their study. First, the lack of precise estimates of the efficacy of the comparators, the utilities of various life states and the absolute short-term risk of thromboembolic complications in mechanical heart valve patients whose warfarin is interrupted. Second, the desirability of hard end points such as mortality and morbidity which were lacking in the model.
Implications of the study
The authors stated that by demonstrating both the inefficiency of the traditional UFH strategy and the benefit offered by the alternative LMWH strategy, they have highlighted the importance of "individualising" the choice of the thromboembolism prophylaxis strategy for each patient. They advocate that those charged with recommending a periprocedural anticoagulation strategy for patients with mechanical heart valves should consider the economic consequences of the UFH approach, in addition to patient safety and medico-legal concerns.
Source of funding
Supported by DuPont Pharmaceuticals, Pharmion, Pharmacia Upjohn and Sanofi-Aventis.
Garcia D A, Libby E N, Rich J S. Perioperative anticoagulation for patients with mechanical heart valves: a model comparing unfractionated and low-molecular-weight heparin. Journal of Clinical Outcomes Management 2005; 12(1): 25-31
Subject indexing assigned by CRD
Anticoagulants /therapeutic use; Cost-Benefit Analysis; Heart Valve Prosthesis; Heparin, Low-Molecular-Weight /therapeutic use; Humans; Perioperative Care /methods
Date bibliographic record published
Date abstract record published