LEADIN G ARTI C LE
PharmaroEc(lnom l('$ )
(~):
417-44 S. ' 99)
,, 7().7690t9}~17/S000fO
C
Ad i~ [ nl~r ... t;" ... [
limitt'd. All n p llS tnrn-nl.
P£Cl 22&
Novel Monoclonal Antiendotoxin Antibody Therapy
Efficacy at Any Price? William K. Fal/t
College of Pharmacy. University or Cincinnati. Cincinnati. Ohio. USA
The United States spends in excess of$US2 billion each da y for healthcare. This expenditure. approaching 12% of the Gross National Product. represents the largest per capita heaJthcare cost in the world. The future direction of health care in the US. and mechanisms fo r providing universal healthcare coverage, were debated in the recent current presidential election (Angell 1992; Clinton 1992; Enthoven 1992; Reinhardt 1992; Sullivan 1992). In the absence oCa national plan Cor defini ng and distributing healthcare resources. the cost of healthcare has exceeded the rate of inflation for a number of years. The conflict between our desire for the latest te<:hnology or drug therapy to treat specific diseases, and our unwillingness or inability 10 pay for such therapy, will not be resolved in the immcdiate future (Coi le 1990). In the absence of consensus regarding funding priorities. the decision to use or not to use a new therapy is made at the patient's bedside. This is in contrast to national healthcare programmes avai lable in some European countries and Canada. The effect of new technologies on h ealthcare costs and on the nation's ' health' was re<:ently discussed (Massaro 1990). Since unlimited funding docs not ensu re maximal therapeutic outcome, an appropriate balance between the cost of therapy and the maximum treatment benefit achieved must be obtained to ensure cost-effccti ve healthcarc. Recently, several publications have reviewed the principles of cost effectiveness as well as the cost effectiveness of accepted interventions (Chatterjee
& Srebro 1990; Detsky & NagJie 1992; Freund & Dittus 1992; Hurley 1990; Puma & Lawlor 1990; Schulman et al. 199 1; Smith 1990). In at least one publication. the increase in healthcare costs. associated with the introduction of new technology, has nOI resulted in improvements in patient outcome (Shy et al. 1990). Although complete information regarding t he benefits and costs of new therapies should be provided to the clin ician prior to their introduction, the majority of cli nical trials provide little information to the clinician about mechanisms fo r integrating t he results of the trial into patient care policies (Hawkins 1984). The development of monoclonal a ntiendotox in antibodies for the treatment of serious Gram-negati ve infections has initiated considerable debate regarding the appropriate use of high-cost pharmaceutical products (Greenman et al. 1991 , 1992; McCabe 1992; Spalter 1992: van Deventer 1992; Warren et al. 992: 1 Wenzel 1992; Ziegler & Smith 1992; Ziegler et al. 199Ia.b). Prior to the introduction of these products, most n ewdrug products were replacements for existing therapy and were often associated with only minimal increases in treatment costs. The advantages of the new products. including improved tolerability profiles or better patient compliance, were used by pharmaceutical companies to justify their increased costs. The antiendotoxin antibodies represent a enw group of products that have the potential to alter disease processes without ofTering substantial cost savings.
438
More imponanll y, they will not replace standard drug therapy.
Sch ulman and colleagues have estimated that the total increase in health costs associated with the use of anliendotoxin antibodies may exceed $US2.3 b illion ($US1.5 bi llion in drug com) with strict guidelines for use, and cou ld exceed $US6.2 billion if their use is unreSlficlcd (Sch ulman et al. 1991). Although cost has not been the major focus of recent leIters and ed itorials, the efficacy of these agents has been scrutinised in leading medical journals (Greenman & Schein [992; Greenman e1 a t 1991; McCabe 1992; $pallcr 1992; van DcvenleT 1992: Warren el al. 1992; Wcnzcll992; Ziegler & Sm ith 1992; Ziegler e1 al. 1991a,b). This ankle reviews the processes tradit ionall y used in the US for evaluati ng new drug therapies, a nd mechanisms for contro lling their use into clinical practice. It also attem pts to point o ut limitations that may appl y to the eval uation of these new agents, and it concl udes with recommendations for future evaluation of the cost effectiveness of new products.
1. Monoclonal Antiendotoxin Antibodies and Sepsis The monoclonal antiendotoxin amibodies have recently been evaluated for the treatment of patients with serious G ram-ncgat ive infections (Greenman et al. 199 1; Ziegler et al. 199 1a). The organisms most common ly associated with these infections contain a lipopolysaccharide (endotox in) componem in their cell wall which has been associated with the adverse systemic reactions seen in these patients. This syndrome is com monly referred to as the sepsi s syndrome and when it occurs in the presence of hypotension is called septic shock (Parrillo et al. 1990). The systemic response 10 endotoxin release may result in hypoperfusion and/or dysfunction of major organ systems including the lung, kidney, heart, liver and brain (Bone et al. 1989). The mortality of patients with sepsis syndrome approaches 50%. despite the introduction of antibiotic therapy and cardiovascular support (Par. rillo et al. 1990). Unfortunately, prospective eri·
PharmacoEcO/wmics 3 (6) 1993
teria for the diagnosis of sepsis do ont adequately identify those patients likely to develop sepsis, septic shock or their associated morbidity. In addition, patients with Gram-positive infections may develop sepsis and septic shock, and are unlikely to respond to antiendotoxin antibodies. Although endotoxin has been strongly associated with the development of sepsis, a number of other mediators of the sepsis syndrome have also been identified (Parrillo et al . 1990). A complete description of the sepsis syndrome is provided elsewhere (Bone et al. 1989; Bone 1991 b; Pamllo e t aJ. 1990). Previous work by Ziegler and colleagues demonstrated the ability of a human polyc1onal antiserum to the core region of endotoxin to reduce the mortality associated with Gram-negative infections (Ziegler et al. 1982). This early work subsequently el d to the development of 2 monoclo nal a ntibodies directed at the same target. Nebaeumab (HA- 1A. eentoxin) produced by Ccntocor is a human monoclonal IgM antibody that binds 10 the lipid A core of endotoxin (Teng et a1. 1985). E5 produced by Xoma is a murine monoclonal antiendotoxin antibody with similar binding activity (Harkonen et al. 1988) Recent cl inical trials with these agents demonstrate their ability 10 alter the sepsis process by binding endotoxin , thus reducing the morbidity and mortality associated with serious Gram-negative infections in selected patient groups (Greenman et al. 1991 ; Ziegler et a1. 1991a). However, lack of agreement about the population of patients most likely to benefit from therapy in the published trials with these agents has fuelled considerable debate concerning their use (Bone 1991a,c). Although most authors agree that these products should not be administered to every patient with a cl in ical pict ure of sepsis, general consensus has not been reached with respect to specific guidelines for their use (Bone 199 1c). In addition, detailed information perta ining to the prevention of or reduction in overall morbidity, a potential source of overall cost savings, have not been reported outside of the original publications (Greenman et al. 1991 ; Ziegler et al. 199Ia). Thi s results in large part from absence of disease definitio ns o fsepsis-related morbidity, and to our in-
AntiendOlOxi n Antibodies: Efficacy al Any Price?
ability to predict which patients will proceed on to septic s hock. Since antibiotics and cardiovascular support remain the primary therapy in this patient population, the cost of these new monoclonal products as adjunct therapy will increase the cost of treati ng sepsis (Parrillo et al. 1990). While these agents may reduce or prevent sepsis-associated morbidity and shorten hospitalisation, detailed data describing these issues have not been published (Greenman ct al. 199 1; Ziegler et al. 199Ia). It should be obvious to the reader that from a cost standpoint alone, these agents have attracted the attention of institutions providing treatment for patients with serious infections.
2. Traditional Cost Comparisons and Formulary Decisions Most individuals, evaluating the cost effects of a new trealment regimen, develo p initial estimates of the cost of a new therapy using morbidity and mortality statistics or estimates of disease incidence from either government or primary literature sources. The medical record or phannacy departments arc used to identify patients who received standard therapy for the disease within a specific time period. Using this information, estimates o f the cost of the new therapy are then developed and compared with the costs associated with traditional therapy. Assumptions about efficacy of the agents are usually drawn from published randomised clinical trials comparing the experimental agent to standard therapy. In most instances the expected outcome between the regimens is similar, with slight differences in either potential outcome, dosage regimens, patient tolerance or adverse effect rates. The primary care provider or the patient is then provided with this information and asked to c hoose the therapy offering the grealest benefit at an acceptable COSI. Institutional decisions concerning the appropriate use of a specific agent are made by an institution-based committee, the Pharmacy and Therapeutics (P&T) Committee. This committee was initially charged with maintaining a list of accepted
439
drug products with the goal of ensuring that o nly Quality products were in usc within the i nst itution . In the past, this committee focused primarily on generic or therapeutic s ubstitu tion policies. Increasingly, these comm ittees arc being asked to develop practice guidel ines for the use of specific agents. The purpose is 2-fold: to promote the appropriate usc of drug products and 10 conlain costs. In some cases, the decisions of Ihis com mittee are inf!uenced by competiti ve bidding processes coordinated by TCgional or national buying groups. Clinicians working within the institution are reQuired to work within the established formulary or practice guidelines or to provide wrinen justification for products nOI included in the formulary listing.
3. New Challenges to the Formulary Decision-Making Process In the current era of cost contai nment, expensive new therapies Quickly attract the attention of hospital administrators anempting to contain costs. Most administrators expect the pharmacy department working wilh the P&T Commillee to control the introd uction of new therapies. The goal oflhis process is to control if not to reduce fixed costs. While the costs o f early biotechnology-derived agents were partially offset by the costs o f traditional therapy, the newer agents re present new adjunctive therapies. This may require a shift from personnel budgets to supply budgets as well as budgetary shifts between departments. In addition, the population of patients likely to benefit from treatment continues to grow with each new product class introduced . A brief review of several products introduced during the past decade will highlight the subtle changes seen with each product. The introduction of recombinant human insulin at a price slightly above that of the animal sources was one of the first examples of a replacement strategy with a new biotechnology product. In most institutions, the reduced risk of immune reactions appeared to be wonh the modest increase in product COSI, and subsequently most patients were con-
440
veTted to (he new products by their physicians. The pharmacy department either requested an increase in budget or reduced acquisition costs in other areas 10 cover this additional cost. Epoclin (recombinan t human erythropoietin) is an example of a biotechnology product which represented a new drug therapy with the potential \0 replace other nondrug therapy costs. Epoetin has been shown in a number of studies to reduce the number of blood transfusions required by patients with end-stage renal disease (ESRD) [Stevens Cl al.
1992J. This reduces the risk of blood-borne diseases associa ted wi th transfusion therapy. Although this product may represent a considerable increase in the cost of Ireating ESRD, its use has also been shown to improve the overall quality of life of patients (Stevens et al. 1992). Unfortunately, the improved state of well-being and increased appetite may also increase the time required for complete dialysis. resulting in an increase in the cost of dialysis (Spinowitl e t a1. 1991). Since only the latter cost has been easy to calculate from published clinical studies, thc overall cost effectiveness of this agent has been difficult to determine (Stevens et al. 1992). Although there has been lillie debate about the cost effectiveness of this agent, problems with funding have restricted the use of this agen t in some patients. A redistribution of resources and payments under the Medicare programme are providing some relief for the costs associated with this therapy (for a full review of the pharmacocconomics of epoct in in chronic renal failure see Whittington et al. 1993). Pharmacy departments arc beginning to evaluate the effect of the antiendotoxin an ti bodies on hospital drug costs within the US. In our institution, a 650 bed teaching hospital, we have identified approxi mately 150 patients in a 12-month period with a discharge diagnosis of sepsis. Using enrolment criteria obtained from the 2 published cl inical trials, we estimate that as many as 300 patients may have met the initial criteria for treatment with either agent (Greenman et a!. 1991 ; Ziegler el al. 199Ia). Using a cost of $US3500 per treatment proposed by Schulman ( 1991), this would represent a new drug expenditure of up to
Pharmaco/':conomics J (6) 199J
$USI 050000; an annual increase in the pharmacy drug budget of over 7%. Unlike the situation with previous products, a transfer of funds within the hospital or a reduction in other pharmacy expenditures will be necessary to cover thc cost of antiendOlOxin therapy. Potential sources for this funding can not be determined from the published trials with these agents, and will require individual institutions to identify and evaluate the cost of treating patients with sepsis syndrome. Since it is unlikely that the Healthcare Finance Administration wi ll incrcase the prospecti ve reimbursement payments to cover these costs without additional cost-effectiveness studies, P&T committees will be asked to develop guideli nes to restrict the use of these products to patient popula tio ns most likely to benefit from theiT use. The Infectious Disease Society of America recentl y published a lim ited series of guidelines 10 assist individual practitioners with clinica l decisions regarding the use of these agents (Wenzel et al. \992). An algorithm for patient selection was also recently presented (Schentag 1992). Recently, increased emphasis is being placed on the evaluation of outcomes associated with drug therapy selection and on new responsibilities and opportunities in pharmaceutical care (Hepler & Strand 1991). The role of the P&T Committee and its relationship to quality assurance programmes evaluating outcomes has not been determined in most institutions. Policies developed by the P&T Committee, which are designed to influence patient care, will need to be evaluated prospectively to ensure that they result in the best patient outcomes. Since guidelines and protocols are more difficult to develop than substitution policies, the delay in FDA approval of these products has been applauded privately by some, as a means of delaying these difficult decisions.
4. Conflicting or
Mi~sing
Information
The results of the E5 and nebacumab trials were recently reviewed (Bone 199Ia,c). Although it is difficult to compare the results of the 2 trials directly, because of differences in study design, def-
Antiendotoxin Antibodies: Efficacy at Any Price?
inilions o fshock and methods of data analysis, the differences in outcome may be partiall y explained by severity of illness, age, and dose regimen. One vicw is that until further information becomes available, the clinical use of these agents should be restricted to patients in whom Gram-negative sepsis is strongly suspected, and who meet the enrolment criteria used in the published cl in ical trial for the p rod uct selccted (Bone 919Ic). An early analysis of the cl inical effi cacy of nebacumab in the repeat trials suggested that the efficacy was less than that in the control arm of the study. These data resulted in terminat ion of the cl inical trial and withdrawal of the product from European markets. The impact of these data on trials of ES is not known. In 1992, Warren and colleagues reviewed the 2 original published clinical trials as well as abstract data presented from a second trial with ES (Warren et al. 1992). They recom mended tha t nebacumab remain experimental until additional studies determine that the products increase surviva l from Gram-negati ve sepsis. They were unable to explain the differences in patient populations who benefited from ES ad ministration in the firs t and second ES studies and consequently made no recommendation concerning the use of ES. At least one author has suggested a that com parati ve trial between the 2 agents might be indicated (Bone 199Ic). For the same reasons, changes in sepsis associated morbidities have been equally difficult to determine from the published literature. Since an alteration in the incidence of morbidity may determine the overall cost effecti veness of these agents, they must be given more attention in future trials. Given the limited information avai lable in the cl inical trials, P&T committees must begin to request detailed information about patients previously treated with a diagnosis of sepsis o r Gramnegative infections. Unfortunately, most medical record systems in the US are not capable o f pr oviding detailed historical information about patients with speci fi c diseases, the outcomes associated with various treatments or the costs of therapy. Very little information about patients with a diagnosis
441
of sepsis will be available \0 P&T committees without ex tensi ve chart reviews. Retrospectively obtained chart information must be integrated wi th the results of the publi shed trials, with 2 principal goals; determination of the number of patients meeting enrolment criteria, and development of guidelines for utilisation of these products. This process will be difficuil and labour intensive. The lack of cri tical historical information is c urrently hinderi ng attempts in our institution to define the potential economic impact of these products and the development of guidelines for use. Since most clinical trials a re not designed 10 capture the costs of therapy or long term outcomes, little information concerning the overall economic impact of these products has been forthcoming, with the exception of recent publications by Schulman and colleagues (1 99 1) and Barriere ( 1992). The P&T Comminee is now faced wi th a new challenge: evaluate the effi cacy of a new agent for which there is no direct alternative, o r determine wh ich patient group may benefi t significantly via reduction in mortality or morbidity if the product is used appropriately. Un like previous therapeutic substitution and form ulary decisions, the decision to wi lhhold thcrapy or to restrict a potentially lifesaving treatment to specific patient groups in the absence of good cl inical data is a difficult choicc. This will undoubtedly raise concerns about healthcare rationing and the ethiC!> of this practice. It is unl ikely t hatconsensus regarding the cost-effective usc of these products will be reached prior to final FDA approval of oneor both products.
5. Pressures to Use Monoclonal Antibodies In 199 1, Bone indicated that h e believed that individual physicians would not feel compelled to use monoclonal antibodies in the absence of good clinical dala (Bone 199 Ic). This author disagrees with this opinion and offers the following. Previously publi shed clinical trials with other anti-infective agents evaluating efficacy used a combination of subjecti vc and/or objective assessments to test this hypothesis regarding efficacy. The efficacy
442
assessments were based on whether the patient was cured , improved, or f ailed to respond to the experi menlal drug. Death on thera pywas rarely eval uated . In con trast, the effi cacy of the antiendotoxin antibodies was based on reductio ns i n m orta lity observed in the experi menta l and control (placebo) a rms o f the W s dies. With a potential red uclion i n mortali ty of approximately one-third in selected pat ien ts, I believe that physicians wi ll feci extreme pressure to usc these agcnts i n pali ems suspected of having Gram -negative scpsis. T his assumpt ion is supported by the li mited guideli nes from the infect ious Diseases Society of America (Wenzel el al. 1992). In the a bsence of inst itution-speci fi c guidelines fo r use, patients who meet publi shed guidelines will be treated wi th antiendotox in antibodies. This will oceur despite the d ifferences in o utcome observed in the various trials. T he FDA's request for add it ional efficacy data fo r nebacumab may provide addi tio nal insight into the popu lation of patients most likely to benefit from therapy wi th these agents.
6. External Assistance K'ith Decision to Use these Products The Federal Government has been watch ing the d evelopment o f n ew products fo r several years. Pro posed Medicare reg ulations w hich w ould have required the Healthcare Finance Administration to approve the cost effectiveness of these new products have been published (Federal Register 1989). However, fi nal approval of these or modified guidelincs is not expected in the immed iate future.
7. Recommendations for Changes in the Drug Development Process In 1990, Beck outlined a process that should be used in the evaluation of new drug pro ducts (Beck 1990). Th is author recommended that the etTeets o f various treatment interventions be modelcd prior to initiating clinical trials. in an attempt to define thera peutic end points for treatment adequately, which is a problem with the current antiendotoxin trials. The importa nce of end- point selection early
PhOfmGcoEronomics
J (6) 199J
in the design of clinical trials has also been reviewed by Terrin (1990). Appropriate end-points will ensure appro priate trial design a nd assist in determi ning the relative meri ts of alternati ve treatments. Beck also recommended that the results obta ined fro m clinical trials be evaluated using an external standard of measure, such as dollars spent per unit of gai n (Beck 1990). T his would imply thai a lo ng with efficacy data, detailed cost information be collected t o quant ify the d irect etTects as well as the a dverse effects associated with treatmenl. The failure of most published trials 10 document the benefit of curre nt cli nical trials to fu ture palients was recentl y reviewed (Haw ki ns 1984). Dctsky and colleagues ad vocate the use of costeffectiveness anal ysis to determ ine improvements in healt hcare o utcomes (Dctsky & Naglie 1992). T his a nalysis s hould usc cl in ical terms such as life years extended or m ortality avoided (Detsky & Naglie 1992). A fi nal step i n the process req uires the redistribution of resources necessary to cover the costs associated wit h newcr therapies (Beck 1990). As might be expected , not all authors agree that healthcare resources shou ld be di stri buted based upon evaluations o f cost effectiveness or q ual ityof-li fe models u ntil these models represent societal expectations for m edical care (Barnett 1991; CarrHill 1992; Puma & Lawlo r 1990; Williams 1992).
8. Cost-Effectilleness Analysis of Nebacumab Schulman a nd colleagues (199 1) evaluated the cost effectiveness of n ebacumab using clinical data repon ed in the publi shed clin ical trial by Ziegler and associates. Using a n an nual discount rate o f 5%, these authors performed a cost-effecti veness ana lysis based solely on direct m edical expenses, Since a difference in outcome was demonstrated only in patients with G ram-negative bacteraemia . the analysis was limited t o this subpopulation. Using estimates of li fe expectancy a nd costs of treatment obtained from a ismilar population of patients in another institution, the authors calculated a costeffecti veness ratio per disco unied year o flife saved of $US24 lOll Ailhough il was not expressed in
AnliendOIOllin Anlibodics: Effic3C)' al Any Price?
terms of qual ity o f ilfe years, this ratio compares favourably wilh that obtained for other thcrapies for which cost-effectiveness analyses have been performed [neonatal intensive care for 500 to 999g neonates (SUS II 055; Boyle et al. 1983); low osmolar contrast media for high-risk patients (SUS22 600; Goel ct al. 1989]. In contrast 10 cholesterol reduction with cotestipol ($US9 1 2S0/year of life gained), th is therapy appeared, in the o pinions of Ihe authors. to be econom ically justified (Kinosia n & Eisenberg 1988). Early identification of patients likely to respond to therapy, using a hypothetical diagnostic test, reduced Ihe ratio to $USI4 900. As expected, changes in the estimated life expectancy of the patien ts were more likely to affect thc cosHffectiveness ralio Ihan either the length of hospitalisation or cost of drug therapy. Unfortu nately, because of limitat ions of the published elinical trials, little information is available concerning long term patient survival. This information is critical before additional cost studies are conducted. The costs associated with reductions in sepsis-induced morbidities were also not included in this analysis. A reduction in sepsisinduced morbidities and subsequently a reduction in the costS of treating these com plications may ultimately provide the cost justi fi cation for the use of these products. In a recent issue of this Journal, Barriere presented the results ofa simi lar study using the direct institutional costs associated with the treatment of 266 patients with Gram-negative sepsis (Barriere (992). A subset of patients (n '" 55) with Gramnegati ve septic shock were chosen for cost analysis, since the g reatest reductions in mortality from the usc of nebacumab had previously been demonstrated by Ziegler and colleagues. In contrast to Schulman. mortali ty rates from the Ziegler publication of 33 and 57% were chosen for nebacumab and placebo, respectively (Ziegler & Smith 1992; Ziegler etal. 199Ia,b). The cost of nebacumab was set al $US3750 per treatment. Cost-effectiveness ratios were calculated for 55 patients wi th Gramnegative septic shock, 2 11 patients with Gram-negative septicaemia and 403 septic patients without regard to aetiology. The cost per life saved was
44]
$US28 950 for patients with Gra m-negative septic shock and $US 115 178 for patients with Gram-negative septicaemia. The cost per life saved in treating all patients with septicaemia was $US205 178. The results o f thi s evaluation in patie nts with Gram-negati ve sepsis surviving for 5 years arc almost identica l to those obtained by Schulman ( 199 1). Using the data obtained from the55 patients with septic shock, Barriere conducted a cosl-benefit analysis using direct expendi tures and data from patients who survived septic s hock without nebacumab therapy. A reduction in hospital stay of at least 4 days was required to justify Ihe cost of nebacumab therapy in this analysis. Unfortu nately. this study did not include the potential for lost reimbursement resulting from reductions in morbidity. These reimbursement cha nges should be evaluated in future studies. Chalfin and colleagues also evaluated the cost effectiveness of monoclonal antibodies in treatment of Gram-negative sepsis. Using average diagnostic accuracy and mortality data from thc published trials of nebacumab and E5, and hospital charges for 1405 patients treated in a tertiary care hospital , the authors developed a treatment algorithm for patients nOl in shock. Although their study demonstrated the potential cost effectiveness of antiendotoxin antibodies, it did not include the costs associaled with changes in sepsis-associated morbidity.
9, Future Directions Decisions regarding drug use within an institution have traditionall y been made by the P&T Committee afler an extensive review of the published clinical d ata. It is unlikely that this process will change in the ncar future. Unlike previous drug compounds, the newer biotechnology products will require the Comm ittee to integrate the results of published trials with data collected within the institution, and to develop institution-specific guidelines for use. Although published guidelines wi ll be of some value, they may not represent the needs of the institution or patient population. This knowledge-build ing process will require access to clinical
444
as well as cost data thaI is c urrently not read ily avai lable fro m most hospital informatio n systems. II will require, al a minimum , the abilit y to identify patients previously treated for specific di seases and to develop institution-specific baseline da ta for comparison. The data on costs must include d irect as well as indirect m edica l expenditures. It should also incl ude inpatient as well as out patient therapy costs. In Ihe future. it is likely that payers including the Federal Government will req uire extensive costeffccti veness studies for all new therapies prior to their adoptio n. As some authors ha ve ind icated . new technologies will be evaluated in term s of their costs and the benefit s received . This wi ll require add itional research into thc process as well as the mechanisms by which cost-effectiveness studies should be conducted. In light o r a ecent r article highlighting our lack or understanding or terminology is this area, considerable education and training will also be necessary ( Lee & Sanchez 199 I). For the present. it appears that the mo noclonal antiendotoxin antibodies are cost effective when used in the populations most likel y to benefit. However, unti l p rospective tools are available to assist with the identification ofthis population, their utilisation should be based on institution-specific guidel ines ror usc.
Acknowledgement Dr. Fa nl has pan icipated in several confe rences on monoclonal antiendotoxi n antibodies and has received honora ria from companies involved in the development of ES.
References An&ell M. The prnidcn lial cand idate$ and heallh care refo"" . New En&land J()IJrnal of Medicine )27: 8QO.SOI. 1992 Barnell OS. Al!.5e$Sment of quali ty of life. American Journal of Cardiology 67: 41C-44C, 1991 Barriert SL The economic im pact of HA· IA (Centoxin) against endOIOlli n. PharmacoEconomiC$ 2: 4OS-413. 1992 Bcek J R. How 10 evaluate drugs. Cost-effectiveness analysis. Journal of the A meric-dn Medica l Associatio n 264: 83·84. 1990 Bo ne Re. Fisher CJ . Ocmme r TP. Slot m an GJ. Metz C Sep$is A. synd rome: a va lid clinical enlity. Crit ica l Care Medicine 11: 389·393. 1989
PharmaroEronQmics J (6) 1991
Bone RC Acrilical evalua lion of Mew agenu for the treatment of sep$is. Journal of the American Medical An ociation 266: 16g6-1691. 19'11. Bo ne RC The pathogenesis of sepsis. Annals of Intcrnal Medi. cine 115: 451-469. 199 1b Bone RC Monoclonal antibodies to endotoxin: new a llies against scp$is7 Journal of the American Med ical As§OCiation 266: 112S1126. ]99lc Boyl~ MH. Torrana GW. Si ncl~if JC', It Ol'Wood St>. &anomie evaluation of neonatal in tensi"e care of very·lo,,·.birth .welihl infants. Nt'W EnaJand Journal of MediciII<' 308: 1330-13)7. 1983 CalT·Hi ll RA. Anoral;n, rHOU
""
Chauer;ee K. Srebro J P. Should the asymptoma tic- pa tienl with a significant proximal UD stenosis underao PTC"? CardiovaseularClinics 21: )1-41. 1990 Chalfin DB. Hollxin ME. Fein AM. Gra/jano CC Cost-cffcc· tivencs, of monoc lonal antibodies to Gram·negluivc endolox in in the treatmenl ofGram·ncga ti ve sepsis in ICU pa,ienlS. Journal of thc American Medical Associ ation 269: 249.254.
""
Clinton ll. The Oinlon health care plan. Ne ..... England Journal of Medicinc 321: 804-807. ]992 Coik Jr RC Te.:hnotOlYand ~Ihic!l: Ihree scenarios for lilt' 1990s. Quanerly Review Bulletin 16: 102-208. 1990 Detsky AS. Na,lie IG. A cIinician's guide to oos1-cfftt,iveness analrsi$. Annals of I nternal Medici ne I I): 141_1S4. 1992 Em ho"en AC. Commentary: measurinK lhe candidatC1 on health care. New England Journal of Medicine )27: 807.8O'J. 1992 Freund DA. Dillus RS. Pri nci ples of ph anna COttO nomic analysis of drug ,herapy. PharmacoEconomics I: 20-31. 1992 God v. Deber RB. Dcuk y AS. Nonionic con trast media: e.:onomic analysis and healt h pO licy deve lopment. Canadian Medical Association Journa l 140: 389-395. ]989 Gr.-cn ma n R L.Schein RMH . Antibody 10 e n doto~in in the lreat· men t of aram·ncgat;'·c sepsis. Journal of the American Med ica] Association 167: 2326. 1992 Gfffnman RL$chcin RM H. Man in M A.Wen1el R P.Macl nt)'re NR. A controlled clinical trial of £5 muri ne monoclonal I,M antibody to endotox in in the treat ment or&ram·~ti'·e sep$I$. Journal of lhe Ameriean Medical Association 266: 1097. ] 102. 1991 Hart.onen S. Scan non P. Misehak R .Spi ller L E. Foull C', Phase I study of a murine monoclonal anli lipid A antibod)' in bac. teremic and no nbacteremic patien ts. Antimicrobial Agen ts and Chemothel':lpy 32: 110-1 16. 1988 Hawkins BS. Eva lua ling Ihe benefIt of clinical trials to future pa tients. Controlled Clinical Tri~l s 5: 13-32. 1984 Health Care Financing Ad min isll':l tion. US [)( pt of Health and Human ScrviCC$. Medicare Program: criteria and procedures for makinK medica] scrvice covcrag decisions that relate to medicalte('hnotOSY. Federal Register S4: 4)02-4318. 1989 Hepler CD. Straoo LM. Opponun ilie$ aoo rnponsibilitiC1 in ph.arml('('utieal cart. American Journal of Hospital Pharmacy 47: B3-54). 1991 Hurley S. Areview orcosl-elfective~ analy$is. Med ical Journal of Austra]ia IS): S20-523, 1990 Kinosian S P. Eisenberc JM . CUlli ng inlO cholesterol: COSt-effe.:. tive altemative$ for trea tina h)'perc~terolemia. Journal of the American Medi~.1 Associalion 259: 2249·22~4. 1988 l« JT, Sanchel LA. Inlerpreta tion of .COSt-effeclive. and sound· ne~ of economic eval ualions i n Ihe pharmacy litera ture. American J ournal of Hospital Pharmacy 48: 2622_2627. 1991 Ma naro TA. Impact of new tec hnologies on healt h-care COStl and on Ihe nation's .health•. O in ical Chem istry 36: ]612·1616. 1990 McCabe WR. Antibody to endoto~in in the treatment of lram-
AnliendOloxin Antibodies: Efficacy al Any Price?
negali"e sepsis. Journal oftllo: Americ3n Medical Associalion 267; 232 S. 1992 Parrillo JE. Parker MM . Nata nson C". Suffrcdinl AF. Danner RL. el at. SePlie Shoc k in huma ns: advan~~s in I h~ underslandi ng of pa lhOgencsis. ca rd iovascu lar dysfunction. and Ihcr.l PY. An· nals of Inlernal Medicine 113: 227-242. 1990 Puma JL. Lawlor EF. Quality-adjusled l ife-~ears: elhical impli~alions for ph ysicians and policy make".. Journal orthe Amer_ iC1ln MedIcal Association 263: 2917-292 1. 1990 Reinhardt UE. Commen\al1~ polili("$ and 1110: health ClIre system. New England Journal of Medicine 327: 809-811. 1992 Scllo:nta, JJ . Antibody to endotoxin in tllo: treaUTlCnI of gramncpli"e sepsis. Journal of the American MediC1l1 Association 261: 2326-2327.1992 Schulman KA . G lick HA. Rubin H. Eisenbe,.. J M. CosH::ffec_ livent» of HA- IA monoclo nal anlibody for gram-negative sepsis. JOUrnal oflhe American Medical A$$OCiation 266: J466.. 3471. 19<.11 Shy KK . Luthy DA. Bcnncn FC. Whitfield M. Lar:;un ED. Effecls of elcclronic felal-hea n -ra le moni to ring.. as com\»rtd wi th pe_ riodic auscultation. on thc ncurolQ&ic development of premature infants. New England Journal of Medicine 322: 588. 593. 1990 Sm ith GT. The economics of h ypene nsion and stroke. Ameri("lln Hean Joomal 119: 72),727. 1990 Spal ter J. Antlbod~ 10 endotoxin in lhe treatment of gram'lIegali"e sePSIS. Journal orlbe Al"OCri<:2n Medical Auociation 267: 2325. 1992 A".lanian 1. Chal1'\an C. Golden RA . Rasroff J. Spinowi tz GaUer M. Impact of epoetin bela on dial)'ler dearantt and hepa rin requirements. America n Journal of Kidney Discases 18: 66$·613. 1991 Slevens M E. Summerfield G P. Hall AA.. Beck CA. Hard inl1 AJ. el al. CO~1 bcnefilS of low dose subculaneous crylhropoiet in in palients with anaemia of end stage renal d isca:;c. Bril ish Medi . cal Journal 3-04: 474.471.1992 Su ll ivan LW. The Bush ad ministrallon'S health care plan. Ne w EngJand JOurnal of Medicine 327: 80 1-804. 1992 Teng NNII. Kaplan IlS. Heben JM. Moon: C. DougJas II . et al.
as.
'" Protection against lram· ....g:n ive bacten:mia and endotoAemia wi th human monoclonal 111M anlibodies. I'roc~wings of the National Academy of Sciences oflhe USA 82: I 19().1 794. 1985 TCrTin ML Efficient uSC of end poin ts in clinical trials: H perspecti ve. Stati"ics in Medicine 9: 515- 160. 1990 van Dcventer S. Antibody 10 end otoxin in the trea lment or,ram. ncgalive sepsis. Journal of the American Medical Assoo:iation 267: 232~2326. 1992 WarTen HS. Dan ner RL. MunfOrd RS. An t i~ndotoxi n monocloru.1 antibodies. New England Journal of Medicine 326.: J IHliS}. 1992 Wenzel RP. Anti~ndOloxin monoclonal amibodics-
COlTCspondence and reprinls: Dr William K. Fum. College of Pharmacy. Universi ty of Cincinnati. 3223 Eden Ave .. Cincinnati. O H 45261-Q004. USA.