Original article The cost of antiretroviral drug resistance in HIV positive patients

Transcription

1 This pdf is for personal use only. To obtain commercial reprints, Antiviral Therapy: in press (doi: /IMP2709) Original article The cost of antiretroviral drug resistance in HIV positive patients Hartmut B Krentz 1,2 *, Karen Ko 1, Brenda Beckthold 1, M John Gill 1,2 1 Southern Alberta Clinic, Calgary, AB, Canada 2 Department of Medicine, University of Calgary, Calgary, AB, Canada *Corresponding author hartmut.krentz@albertahealthservices.ca Introduction: Antiretroviral (ARV) drug resistance limits treatment choices, often necessitating the selection of drugs with less desirable pill burdens, toxicity profiles, drug interactions and dosing schedules, and may increase cost. We examine the impact of resistance on the cost of HIV care. Methods: All adult HIV-positive individuals newly referred for HIV care from 1 January 2007 to 1 January 2011 and followed until 31 December 2011 at the Southern Alberta Clinic, Calgary, AB, Canada, were included. We determined the cost of all ARV drugs, and HIV-related outpatient and inpatient care, reported as mean per patient per month (PPPM) costs in 2011 Canadian dollars (CDN). Results: Overall, 78% of patients received genotypic antiretroviral resistance testing (GART); 57% among ARV-naive patients, 21% after suspected viral failure, and 20% during a treatment interruption. Resistance was detected in 6%, 49% and 21% of all tests respectively. The total mean PPPM cost for patients with GART was CDN 1,179. Patients with primary resistance had mean total PPPM costs of CDN 956. Patients with no resistance had mean PPPM costs of CDN 1,058, by contrast with the CDN 1,291 costs of patients with secondary/acquired resistance. Mean costs for one, two or three ARV class resistance was CDN 1,278, 1,337 and 1,801, respectively. Mean PPPM costs increased by 31% from CDN 1,068 to 1,396, respectively, before and after a positive resistance test. Conclusions: Transmission of resistant virus as well as emergence of resistance during ARV therapy leads to increased costs. Mean costs increased by number of resistant classes. Routine GART and optimizing ARV regimens to avoid resistance might minimize the long-term costs of HIV care. Introduction Genotypic antiretroviral resistance testing (GART) has become a cornerstone in HIV care delivery and is recommended in published guidelines from both the US [1] and Europe [2,3]. Resistance testing was initially introduced for patients failing antiretroviral (ARV) therapy (that is, secondary resistance) to define which drugs or classes of drugs were unlikely to be contributing to the control of HIV replication. More recently, GART use was expanded to include testing ARV drugnaive patients starting treatment so as to avoid use of ARV where underlying polymorphisms or primary (that is, transmitted) drug resistance patterns would likely make them ineffective [4 6]. While the current population-based GART technologies may be imperfect, occasionally missing resistance due to sampling issues, the presence of documented resistance still has major clinical implications [7 13]. Resistance limits the choice of ARV options for the prescribing physician. Restriction of choice may lead to selection of drugs or regimens with less desirable pill burden, toxicity profiles, drug interactions and dosing schedules, and often at an increased cost due to the need to use newer more expensive treatment regimens. The ongoing accrual of new resistance mutations during any subsequent failing treatment regimens further exacerbates these issues leading to further limitations of choices. While the impact of resistance to one or more classes of ARVs has been shown to be clinically important by leading to worse medical outcomes [14], the impact on the cost of care from primary resistance in ARV-naive patients or on secondary resistance in patients failing ARV regimens has not been measured. Such information could be useful in evaluating many economic aspects of HIV care such as the cost justification of current and future more sensitive resistance testing methodologies, the relative economic value of low and high genetic 2014 International Medical Press (print) (online) Page numbers not for citation purposes 1

2 HB Krentz et al. This pdf is for personal use only. To obtain commercial reprints, barrier drugs or regimens thereby influencing drug selection, the cost justification of adherence enhancing measures, and, in the developing world, the justification of optimizing inventory size to minimize risk of supply interruptions leading to treatment failures. We wished to measure the direct costs of HIV-related care (that is, ARV drugs, HIV-related inpatient hospitalizations and HIV-related outpatient care) in all HIVpositive patients, and to examine specifically the costs for patients with GART who had either no resistance or had documented resistance mutations (either primary or secondary). We also wished to compare costs before and after a resistance mutation appeared, and to examine the cost impact of one-, two-, or three-class (that is, nucleoside reverse transcriptase inhibitor [NRTI], non- NRTI [NNRTI] and/or protease inhibitor [PI]) resistance to ARV drugs. Methods Study population All HIV-positive individuals referred for HIV assessment and initiating care are automatically enrolled in the Southern Alberta Cohort (SAC). SAC includes all patients living anywhere in Southern Alberta who receive or have received HIV care through the regional HIV care programme located in Calgary. Access to provincially funded HIV care, including HIV-specific laboratory tests and free ARV therapy is provided exclusively through this programme. For this study all adult ( 16 years) HIV-infected patients initiating HIV care within SAC either due to a newly diagnosed HIV infection or having been diagnosed elsewhere and moving into the region from 1 January 2007 to 1 January 2011 were included and followed until they died, moved, were lost to follow-up (LTFU), or until 31 December LTFU is defined as having no clinic contact for 12 months including no lab work or drug refills. To ensure adequate follow-up, patients were included if they had 180 consecutive days of follow-up from their initial SAC visit. We placed patients into the following comparative categories as shown in Figure 1: patients who never had GART, patients with GART but no documented resistance, patients with primary resistance (ARV-naive at time of testing), and patients with secondary resistance. Data collection Routine administrative, demographic and clinical data were obtained retrospectively for all patients from the SAC database. Demographic data including sex, age at diagnosis, self-reported ethnicity (Caucasian, Aboriginal/Native Canadian, Black and Other), location of HIV diagnosis and initial HIV treatment (that is, locally diagnosed in southern Alberta/outside of southern Alberta) Figure 1. Study population: 1 January 2007 to 1 January 2011 All patients accessing HIV care at SAC (n=601) No GART (n=130) No resistance (n=401) GART (n=471) Secondary resistance (n=49) Primary resistance (n=21) GART, genotypic antiretroviral resistance testing; SAC, Southern Alberta Cohort. and HIV risk factor were determined at initial clinic visit. HIV risk factor was defined as men having sex with men (MSM), men having sex with women (MSW), intravenous drug use (IVDU), and other (for example, blood transfusion). Clinical data including CD4 + T-cell count at HIV diagnosis, CD4 + T-cell count at initial SAC visit, CD4 + T-cell count at positive resistance test, AIDS defining illnesses and comorbidities were determined at the initial visit and updated at subsequent clinic visits. Genotypic antiretroviral resistance testing Overall, 95% of GART tests (VircoTYPE HIV-1; Janssen Diagnostics BVBA, Mechelen, Belgium) were drawn while the patients were followed at SAC. Resistance to a class was defined by the VircoTYPE HIV-1 analysis showing 1 agent as resistant or minimal response. Once class resistance was documented it was deemed irreversible. Reasons for testing were categorized as suspected viral failure, treatment interruption, ARV-naive initiating ARV treatment, and other. We examined drug resistances to three major classes of ARV drugs: NNRTIs, NRTIs and PIs. Although both a fusion inhibitor and two integrase inhibitors were used during the study period, we only found one instance of documented integrase inhibitor genotypic resistance. Cost of HIV care HIV care was defined as ARV drugs, HIV-related outpatient visits to SAC, and all HIV-related hospitalizations. Direct costs of care were collected and analysed using a methodology previously described [15]. Basically, ARV drug costs, lab utilization, and outpatient care costs were derived directly from the SAC database whereas inpatient costs (that is, unit service costs) were supplied by the regional health service providers. Costs associated with outpatient care included all HIV clinic visits, visits to HIV-related specialists, visits to other physicians, and all laboratory testing. 2 Page numbers not for citation purposes 2014 International Medical Press

3 This pdf is for personal use only. To obtain commercial reprints, The cost of ARV resistance Laboratory test costs include CD4 + T-cell count, viral load, GART, serological tests, haematology and routine chemistry testing. Costs were calculated as cost per test per patient. All HIV-related inpatient care costs were obtained from the Alberta Health Services Data Integration, Measurement and Reporting Department (Calgary, AB, Canada). These costs were calculated on a per specific patient per day basis based upon the length of the hospital admission and included both the direct cost associated with patient care (that is, drugs, supplies, equipment, salaries, nursing care, laboratory and diagnostic testing) and indirect costs associated with hospital overhead (that is, administration and support services). To confirm correct coding of diagnosis (based on ICD-9 and ICD-10 codes), one author (MJG) also reviewed the hospital discharge summaries of all SAC patients to verify if they were HIV- or non-hiv-related admissions. The unit costs used are market values charged to the regional payer (Calgary Health Region). In order to differentiate inpatient costs occurring at HIV diagnosis versus costs that may be associated with ARV resistance, we only included inpatient costs >30 days after HIV diagnosis. All costs were obtained directly from the costing agencies and reported in Canadian dollars (CDN) adjusted for inflation to 2011 (during the study period CDN 1 equalled approximately 1.02 US dollars). For comparative purposes we used mean cost per-person, per-month (PPPM) determined by multiplying service usage by unit service cost divided by the total number of SAC patients followed during the year. We used a net costing approach [16 18] initially to compare costs between patients who had or had not received GART from the healthcare provider (that is, Alberta Health Services) perspective. We then compared costs of patients who had GART but no resistance to patients demonstrating either primary or secondary resistance. For patients with secondary resistance, we compared costs before the date of resistance from their SAC start date to the date of the test and then from this date forward until they moved, were LTFU, died or 31 December If a patient had >1 positive resistance test, we use the date of the first test. We determined the cost of care of patients with one, two or three ARV class resistances. Patients were also stratified into three CD4 + T-cell categories based on their initial CD4 + T-cell count: <200 cells/mm 3, cells/mm 3 and >350 cells/mm 3. Initial CD4 + T-cell count rather than current or most recent CD4 + T-cell count was used as it better reflects cost of care in the long-term controlling for CD4 + T-cell rebound [19]. In order to assess ARVrelated costs, patients were subdivided into two treatment groups: ARV-naive or ARV-experienced (that is, currently on or ever on ARV treatment in the duration of the study period). In order to determine if length of follow-up time affects costs, we conducted a subanalysis comparing patients at 6 month intervals of follow-up. Ethics Patients accessing care and treatment at the Southern Alberta Clinic (SAC), Calgary, AB, Canada, voluntarily sign a written informed consent form authorizing the use of confidential administrative data for research purposes in a manner approved by ethics. The use of non-nominal data of the SAC cohort for research has been approved by the University of Calgary Conjoint Health Research Ethics Board. Statistical analyses Nominal regression analyses were performed for the variables no GART versus GART, and for resistance testing results: secondary resistance versus no resistance and primary resistance versus no resistance. In each case, adjustments were made for age (continuous), gender (male versus female), initial CD4 + T-cell count (continuous) and HIV risk factor (MSM, MSW, IVDU or other). Statistical significance was defined as P<0.05 and 95% CIs were reported. Statistical analyses were performed with SPSS version 20.0 (IBM, Armonk, NY, USA). Results Overall, 78% (n=471) of the 601 patients starting HIV care at SAC from 1 January 2007 to 1 January 2011 and followed for >180 days received genotypic HIV-1 resistance testing. Patients not receiving GART (22%), when compared with patients receiving GART, were older (39 versus 37 years), non-caucasian (60% versus 48%), had MSW as the most likely HIV risk factor (63% versus 49%) and were more often diagnosed outside of southern Alberta (73% versus 33%; all P<0.01; Table 1). There were no differences in initial CD4 + T-cell count or having an AIDS diagnosis at HIV diagnosis, however patients not tested were much more likely to have been living longer with an HIV diagnosis (56 versus 2 months; P<0.001); patients living longer with HIV were more likely to have been diagnosed elsewhere before moving to southern Alberta (P<0.001), and more likely to be virally suppressed at their initial SAC clinic visit (P<0.05) compared to more recently diagnosed patients (data not shown). Since 1 January 2007, 58.7% of all GART (n=1,037) were performed for ARV-naive patients beginning or considering ARV therapy, 22.1% after a suspected viral load failure on ARV therapy, and 20% during an ARV treatment interruption. Some patients experienced >1 episode of viral failure and/or treatment interruptions resulting in multiple GART. Antiviral Therapy Page numbers not for citation purposes 3

4 HB Krentz et al. This pdf is for personal use only. To obtain commercial reprints, Table 1. Sociodemographic and clinical variables for all HIV-positive patients newly accessing HIV care at the Southern Alberta Clinic from 1 January 2007 to 1 January 2011 Variable Total patients No GART With GART No resistance Resistance Primary Secondary Patients, n (%) 601 (100) 130 (21.7) 471 (78.3) 401 (66.7) 70 (11.6) 21 (3.5) 49 (8.2) Gender Male, n (%) 424 (70.5) 91 (21.5) 333 (78.5) 279 (65.8) 54 (12.7) 17 (4.0) 37 (8.7) Female, n (%) 177 (29.5) 39 (22.0) 138 (78.0) 122 (68.9) 16 (9.0) 4 (2.3) 12 (6.8) Median age, years (IQR) a 38 (31 45) 39 (34 46) b 37 (30 45) 37 (30 45) c 39 (33 45) 39 (31 45) 39 (31 45) Self-reported ethnicity Caucasian, n (%) 296 (49.3) 50 (16.9) b 246 (83.1) 208 (70.3) 38 (12.8) 15 (5.1) 23 (7.8) Non-Caucasian, n (%) d 302 (50.7) 78 (25.8) 224 (74.1) 192 (63.5) 32 (10.5) 6 (2.0) 26 (8.6) HIV risk factor MSM, n (%) 227 (37.8) 37 (16.3) 190 (83.7) 165 (72.3) 25 (11.0) 9 (4.0) 16 (7.0) MSW, n (%) 318 (52.9) 83 (26.1) b 235 (73.9) 200 (62.9) 35 (11.0) 10 (3.1) 25 (7.9) IVDU, n (%) 50 (8.3) 10 (20.0) 40 (80.0) 31 (62.0) 9 (18.0) 2 (4.0) 7 (14.0) Other, n (%) 6 (1.0) 0 6 (1.0) 5 (83.3) 1 (16.7) 0 1 (16.7) HIV diagnosis location e Local, n (%) 340 (56.6) 24 (7.1) 316 (92.9) 281 (82.6) 35 (10.3) 14 (4.1) 21 (9.1) Non-local, n (%) 250 (41.6) 95 (38.1) b 155 (61.8) 119 (47.8) 35 (14.0) c 0 (0.0) 35 (14.1) Not reported, n (%) 11 (1.8) 3 (27.3) 8 (72.7) 8 (72.7) Median CD4 + T-cell count at 302 ( ) 293 ( ) 306 ( ) 319 ( ) 231 (97 360) c 187 (59 270) 235 (97 497) HIV diagnosis, cells/mm 3 (IQR) AIDS at HIV diagnosis, % Median time since HIV 4 (1 53) 56 (12 90) b 2 (1 29) 2 (1 21) 12 (1 82) c 1 (1 2) 56 (4 105) diagnosis, months (IQR) f Only patients with >180 days of follow-up are included. Includes patients either with a new diagnosis or transferring from outside of the region. a Age at initiating HIV care at the Southern Alberta Clinic, Calgary, AB, Canada. b P<0.01 for comparison between patients with versus without genotypic antiretroviral resistance testing (GART). c P<0.01 for comparison between patients with GART with versus without primary or secondary resistance. d Non-Caucasian ethnicity includes Aboriginal/Native Canadian, Black, Asian, Hispanic and other. e Local refers to HIV diagnosed and initially treated in southern Alberta and non-local refers to HIV diagnosed and initially treated outside of southern Alberta. f Time since HIV diagnosis from HIV-positive date to the initial clinical visit at the Southern Alberta Clinic. IVDU, intravenous drug use; MSM, men having sex with men; MSW, men having sex with women. Overall GART demonstrated resistance in 70 (14.9%) patients tested; 21 (4.5%) patients had primary resistance, whereas 49 (10.4%) had secondary resistance. A total of 19 (90.4%) patients with primary resistance had resistance to one class of ARVs (15 NNRTI, 2 NRTI and 2 PI); 2 (9.6%) had resistance to two classes (NRTI and NNRTI), while no three class resistances were documented. By contrast, 22 (44.6%) patients with secondary resistance had mutations to one class (16 NNRTI and 6 NRTI), 20 (40.8%) to two classes (NNRTI and NRTI) and 7 (14.3%) to three classes. Resistance was most often found in the context of viral load failure (49%) or after a treatment interruption (21%). Patients with any resistance were more likely to be older (39 versus 37 years), non-local (23% versus 12%), to have had a lower median initial CD4 + T-cell count (231 cells/mm 3 versus 319 cells/mm 3 ) and to have lived longer with documented HIV infection at the time of GART testing (median 12 versus 2 months; all P<0.001) than patients with no resistance, as shown in Table 1. There were no sociodemographic or clinical differences between patients with primary or secondary resistance. Cost comparisons Overall, 97 (16.1%) patients initiating care at SAC did not initiate ARV therapy during the study period and were excluded from the costing analysis. Cost comparisons for the remaining 504 (83.9%) patients are listed in Table 2. The mean ±sd total cost of HIV care (PPPM) for all ARV-experienced patients was CDN 1,109 ±276; 76% (CDN 848) consisted of the cost of ARV drugs, 23% (CDN 250) for outpatient costs (that is, clinic visits and laboratory tests), and 1% (CDN 11) for HIV-related inpatient hospitalizations. Total costs, ARV drug costs and inpatient costs all increased with decreasing initial CD4 + T-cell counts (P<0.05). Outpatient costs were similar for all CD4 + T-cell categories. Mean costs did not vary significantly by number of months of follow-up care. Patients who did not receive GART had higher total costs (CDN 1,192 ±246 versus 1,083 ±304) PPPM (P<0.001) with higher monthly ARV costs (CDN 983 versus 807) but lower outpatient costs (CDN 209 versus 262), respectively, compared to all patients who received GART. Costs were proportionately higher for all initial CD4 + T-cell categories. 4 Page numbers not for citation purposes 2014 International Medical Press

5 This pdf is for personal use only. To obtain commercial reprints, The cost of ARV resistance Table 2. Mean cost PPPM for HIV care in for new patients initiating care at the Southern Alberta Clinic, Calgary, AB, Canada, between 1 January 2007 and 1 January 2011 Patients by category Patients, n Total costs PPPM ARV drugs PPPM Outpatient PPPM Inpatient PPPM All ARV-experienced patients 504 1,109 (276) 848 (259) 250 (82) 11 (80) 200 cells/mm ,213 (284) 940 (261) 256 (85) 18 (105) cells/mm ,098 (245) 855 (222) 255 (87) 10 (90) >350 cells/mm ,005 (265) 759 (264) 242 (73) 3 (20) Patients with no GART 122 1,192 (246) a 983 (231) a 209 (56) a 10 (77) 200 cells/mm ,236 (245) 1,022 (244) 214 (49) 10 (77) cells/mm ,203 (226) 992 (211) 211 (51) 0 (0) >350 cells/mm ,165 (259) 950 (234) 214 (54) 17 (56) All patients with GART 382 1,083 (304) 807 (280) 262 (85) 13 (91) 200 cells/mm ,207 (310) 921 (282) 266 (88) 22 (116) cells/mm ,070 (267) 791 (233) 265 (91) 13 (100) >350 cells/mm (281) 683 (276) 254 (74) 4 (23) Patients with no resistance 314 1,058 (281) 783 (253) 260 (83) 14 (99) 200 cells/mm ,179 (253) 891 (223) 263 (86) 24 (13) cells/mm ,074 (284) 795 (239) 266 (91) 14 (111) >350 cells/mm (262) 652 (259) 254 (74) 3 (18) Patients with primary/ transmitted resistance Total (303) b 686 (275) b 265 (89) 6 (26) 200 cells/mm ,039 (374) 791 (317) 248 (90) 0 (0) cells/mm (276) 654 (260) 304 (101) 0 (0) >350 cells/mm (214) 361 (268) 248 (42) 38 (67) Patients with secondary resistance Total 48 1,291 (380) c 1,007 (378) c 270 (92) 14 (53) 200 cells/mm ,492 (378) 1,180 (380) 293 (99) 19 (64) cells/mm ,287 (307) 1,017 (338) 257 (91) 13 (53) >350 cells/mm ,152 (304) 889 (248) 254 (79) 9 (37) HIV care was defined as antiretroviral (ARV) drugs, outpatient visits and HIV-related inpatient admissions. Data reflect mean and (sd) Canadian dollars (CDN) in 2011 unless indicated otherwise. Only ARV-experienced patients receiving ARV therapy are included. CD4 + T-cell counts were reported as initial value after HIV diagnosis. a No genotypic antiretroviral resistance test (GART) versus GART, P< b Primary resistance versus no resistance, P< c Secondary resistance versus no resistance, P< PPPM, per patient per month. ARV-naive patients with primary resistance and who subsequently initiated ARV therapy had lower mean PPPM total costs of HIV care (CDN 956 ±303) compared with ARV-naive patients starting therapy with no resistance (CDN 1,036 ±286; P<0.05) driven mainly by lower ARV drug costs (CDN 688 versus 774, respectively; data not shown). By contrast, ARV-experienced patients who subsequently developed secondary resistance had mean PPPM total costs that were 22% higher than patients without resistance (CDN 1,291 ±380 versus 1,083 ±304; P<0.001). The higher cost for patients with secondary resistance was primarily due to higher mean monthly ARV drug costs (CDN 1,007 versus 807; P<0.001). The higher costs were found across all CD4 + T-cell categories. Outpatient and inpatient costs were not statistically different. Mean PPPM total cost of HIV care increased with increasing number of ARV class resistances a patient was experiencing: CDN 1,278 ±232 for one class, CDN 1,337 ±243 for two class, and CDN 1,801 ±241 for three class resistance (P<0.001). The cost of ARV drugs account for most of the increase (CDN 932 versus 987 versus 1,513, respectively). For patients with a minimum of 90 days follow-up before and after a positive resistance test (n=29), the cost of HIV care increased by 31% after showing new resistance from CDN 1,068 ±221 to 1,396 ±287 PPPM (P<0.001) with ARV, outpatient and inpatient costs increasing from CDN 813, 228 and 27 to CDN 959, 277 and 159, respectively as shown in Figure 2. For the majority of patients, the number of drugs within an ARV regimen increased from three to four contributing to most of the increased costs. Discussion We found that nearly 1 in 7 patients receiving GART had either primary or secondary resistance to one or more classes of ARV drugs. Since the start of 2007, 4.5% of ARV-naive patients had primary resistance. Buchacz et al. [20] and others [21 23] have shown Antiviral Therapy Page numbers not for citation purposes 5

6 HB Krentz et al. This pdf is for personal use only. To obtain commercial reprints, Figure 2. Mean costs PPPM before and after testing positive for ARV resistance: ARV-experienced patients only 1,600 Before GART After GART 1,400 1,200 Mean PPPM, 2011 CDN 1, Inpatient Outpatient Before GART, CDN After GART, CDN ARV drugs Total costs 1,068 1,398 n=29. ARV, antiretroviral; CDN, Canadian dollars; GART, genotypic antiretroviral resistance testing; PPPM, per patient per month. that both the prevalence and incidence of ARV resistance has decreased over the past decade due primarily to changes in HAART regimens that are more effective and tolerable. Initiating HAART at higher CD4 + T-cell levels may also lower the frequency of ARV resistance mutation as discussed by Uy et al. [24]. Resistance was discovered in nearly half of patients tested during a viral load failure but in only 20% of patients after a treatment interruption. This may be a low estimate of the true prevalence of resistance as timing of GART during a treatment interruption may obscure low prevalence but clinically significant viral populations. The presence of resistance mutations increased the cost of HIV care for ARV-experienced patients mainly due to the higher costs of ARV regimens. Overall, mean PPPM costs were 22% higher for patients with secondary resistance mutations. The use of more complex, newer, more expensive ARV regimens with increased pill burden per regimen contributed to higher costs. The cost of drugs was also higher depending upon a patient s initial CD4 + T-cell count, especially for patients with initial CD4 + T-cell count 200 cells/ mm 3, suggestive of aggressive therapy in face of little manoeuvring room. Outpatient costs were slightly higher indicating more monitoring of the patient s condition although inpatient costs were variable. In a previous study [15], we found that inpatients costs have remained relatively low (that is, as a percentage of total HIV care costs within a population) and increase with decreasing CD4 + T-cell count; here, we find similar trends. Resistance does not appear to differentially increase the rate or cost of inpatient care in our study population. Costs of HIV care, however, increased significantly before and after the discovery of resistance. Not surprisingly, costs were higher for patients with resistance to >1 class of ARV drugs. ARV-naive patients with primary resistance had equivalent or only slightly lower costs of HIV care than patients with no resistance starting ARV therapy in contrast to the much higher costs for patients, particularly ARV costs, with secondary resistance. This may be due to the many different drug options available for first-line treatments that do not produce significant cost differences. Our results suggest that the prudent selection of an effective initial ARV regimen is crucial for patients with resistance as acquisition of further resistance mutations to other classes will drive costs up in the long term. Our study does have limitations. We examined costs in only one HIV centre, albeit a population-based cohort, we only used population-based genotyping, costs were local drug costs, testing was based on clinical use and 6 Page numbers not for citation purposes 2014 International Medical Press

7 This pdf is for personal use only. To obtain commercial reprints, The cost of ARV resistance may not reflect optimal testing to determine resistance and we did not break down costs by agents within class (for example, first or second generation NNRTI or NRTI agents used for HBV therapy). Sample sizes are relatively small, especially for patients with primary resistance, and made smaller by CD4 + T-cell categorization thus making these cost comparisons suggestive rather than definitive. We did not examine patients diagnosed prior to 2006 with resistance who may not have benefited from recent guidelines. We suspect that prevalence of resistance and ARV costs may be higher in these legacy patients. We also did not examine in detail the relationship between poor adherence and ARV resistance, although the two are closely linked [25,26]. Poor adherence may be the contributing factor in the appearance of secondary resistance and lead to the higher costs that we found in this study. The precise costs of care for HIV infection (PPPM) may vary between countries and even between centres within a country for a variety of reasons including drug, laboratory, clinic and hospitalization costs [27]. While the precise increased costs from ARV resistance may also vary somewhat, our costs of care are widely comparable to others and the proportional increases should be similar across centres. Our study suggests that the presence of documented resistance is strongly associated with increased cost of care. GART testing, selection of optimal customized regimens as well as the use of tools and regimens to minimize selection of resistance may all have a role in keeping the cost of HIV care down by limiting the selection of ARV resistance. Acknowledgements We wish to thank Reed Siemieniuk for his assistance and analysis. Disclosure statement The authors declare no competing interests. References 1. US Department of Health and Human Services. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents (Accessed 7 September 2012.) Available from html/1/adult-and-adolescent-treatment-guidelines/6/drugresistance-testing 2. European AIDS Clinical Society. Guidelines for the clinical management and treatment of HIV infected adults in Europe. Version 6.0 (Updated October Accessed 7 September 2012.) files/pdf%20files/eacsguidelines-v6.0-english.pdf 3. Hirsch MS, Gunthard HF, Schapiro JM, et al. Antiretroviral drug resistance testing in adult HIV-1 infection: 2008 recommendations of an International AIDS Society USA Panel. Clin Infect Dis 2008; 47: Little SJ, Holte S, Routy JP, et al. Antiretroviral-drug resistance among patients recently infected with HIV. N Engl J Med 2002; 347: Novak RM, Chen L, MacArthur RD, et al. Prevalence of antiretroviral drug resistance mutations in chronically HIVinfected, treatment-naive patients: implications for routine resistance screening before initiation of antiretroviral therapy. Clin Infect Dis 2005; 40: Sax PE, Islam R, Walensky RP, et al. Should resistance testing be performed for treatment-naive HIV infected patients? A cost-effectiveness analysis. Clin Infect Dis 2005; 41: Wensing AM, van de Vijver DA, Angarano G, et al. Prevalence of drug-resistant HIV-1 variants in untreated individuals in Europe: implications for clinical management. J Infect Dis 2005; 192: Wegner SA, Wallace MR, Aronson NE, et al. Long-term efficacy of routine access to antiretroviral-resistance testing in HIV type 1-infected patients: results of the clinical efficacy of resistance testing trial. Clin Infect Dis 2004; 38: Bansi L, Smith C, Phillips A, et al. The impact of HIV drug resistance testing on changes to treatment. AIDS 2011; 25: Taylor S, Jayasuriya A, Smit E. Using HIV-1 resistance tests in clinical practice. J Antimicrob Chemother 2009; 64: Elgalib A, Perry M, Abound M, et al. Clinical utility of genotypic resistance tests for HIV-1 infected patients with lowlevel virological failure. J Clin Microbiol 2010; 48: Ceccherini-Silberstein F, Cento V, Calvez V, Perno CF. The use of human immunodeficiency virus resistance tests in clinical practice. Clin Microbiol Infect 2010; 16: Paredes R, Cloetet B. Clinical management of HIV-1 resistance. Antiviral Res 2010; 85: Palella FJ, Jr., Armon C, Buchacz K, et al. The association of HIV susceptibility testing with survival among HIV-infected patients receiving antiretroviral therapy: a cohort study. Ann Intern Med 2009; 151: Krentz HB, Gill MJ. Cost of medical care for HIV infected patients within a regional population from 1997 to HIV Med 2008; 9: Lipscomb J, Yabroff KR, Brown ML, et al. Health care costing: data, methods, current applications. Med Care 2009; 47 Suppl 1:S1 S Barlow WE. Overview of methods to estimate the medical costs of cancer. Med Care 2009; 47 Suppl 1:S33 S Barnett PG. An improved set of standards for finding cost for cost-effectiveness analysis. Med Care 2009; 47 Suppl 1:S82 S Krentz HB, Gill MJ. Despite CD4 cell count rebound the higher initial costs of medical care for care for HIV-infected patients persist 5 years after presentation with CD4 cell counts less than 350. AIDS 2010; 24: Buchacz K, Baker R, Ward DJ, et al. Trends in decline of antiretroviral resistance among ARV-experienced patients in the HIV Outpatient Study: Aids Res Treat 2012; 2012: Gill VS, Lima VD, Zhang W, et al. Improved virological outcomes in British Columbia concomitant with decreasing incidence of HIV type 1 drug resistance detection. Clin Infect Dis 2010; 50: von Wyl V, Yerly S, Boni J, et al. Long-term trends of HIV Type 1 drug resistance prevalence among antiretroviral treatment-experienced patients in Switzerland. Clin Infect Dis 2009; 48: Audelin AM, Lohse N, Obel J, et al. The incidence rate of HIV type-1 drug resistance in patients on antiretroviral therapy: a nationwide population-based Danish cohort study Antivir Ther 2009; 14: Uy J, Armon C, Buchacz K, Wood K, Brooks JT. Initiation of HAART at higher CD4 cell counts is associated with a lower frequency of antiretroviral drug resistance mutation at virologic failure. J Acquir Immune Defic Syndr 2009; 51: Antiviral Therapy Page numbers not for citation purposes 7

8 HB Krentz et al. This pdf is for personal use only. To obtain commercial reprints, 25. Bangsberg DR, Acosta EP, Gupta R, et al. Adherenceresistance relationships for protease and non-nucleoside reverse transcriptase inhibitors explained by virological fitness. AIDS 2006; 20: Bangsberg DR, Kroetz DL, Deeks SG. Adherence-resistance relationships to combination HIV antiretroviral therapy. Curr HIV/AIDS Rep 2007; 4: Accepted 1 November 2013; published online 3 December Beck EJ, Harling G, Gerbase S, DeLay P. The cost of treatment and care for people living with HIV infection: implications of published studies, Curr Opin HIV AIDS 2010; 5: Page numbers not for citation purposes 2014 International Medical Press