Multidrug-resistant tuberculosis (MDR-TB) treatment in the UK: a study of injectable use and toxicity in practice

Transcription

1 J Antimicrob Chemother 2011; 66: doi: /jac/dkr221 Advance Access publication 3 June 2011 Multidrug-resistant tuberculosis (MDR-TB) treatment in the UK: a study of injectable use and toxicity in practice Ann Sturdy 1, Anna Goodman 2, Ricardo J. José 3, Angela Loyse 4, Marie O Donoghue 5, Onn Min Kon 5, Martin J. Dedicoat 6, Thomas S. Harrison 4, Laurence John 2, Marc Lipman 3,7 and Graham S. Cooke 1,5 * 1 Faculty of Medicine, Imperial College London, London SW7 2AZ, UK; 2 Northwick Park Hospital, North West London NHS Trust, London HA1 3UJ, UK; 3 Royal Free Hospital, Royal Free Hampstead NHS Trust, London NW3 2QG, UK; 4 St George s, University of London, London SW17 0RE, UK; 5 St Mary s Hospital, Imperial College Healthcare NHS Trust, London W2 1PG, UK; 6 Heart of England Foundation Trust, Bordesley Green East, Birmingham B9 5SS, UK; 7 Division of Medicine, University College London, London WC1E 6BT, UK *Corresponding author. Faculty of Medicine, Imperial College London, Jefferiss Research Trust Laboratories, 4th Floor, Medical School Building, St Mary s Campus, Norfolk Place, Paddington, London W2 1PG, UK. Tel: ; Fax: ; g.cooke@imperial.ac.uk Received 14 February 2011; returned 3 April 2011; revised 27 April 2011; accepted 6 May 2011 Background: Multidrug-resistant tuberculosis (MDR-TB) is an increasing challenge to health services globally. Although new drugs are in development, current guidelines still recommend prolonged use of injectable antimicrobials (usually amikacin, kanamycin or capreomycin). The evidence base to inform treatment and monitoring strategies is very limited. Methods: We conducted a retrospective study of patients initiating injectable antimicrobials for MDR-TB treatment in five UK centres between January 2004 and December (i) Current treatment and monitoring strategies were reviewed. (ii) The incidence of ototoxicity (defined both clinically and on audiological testing) and factors associated with ototoxicity were investigated using logistic regression. Results: (i) The choice of injectable antimicrobial varied. Of 50 MDR-TB patients, 29/50 (58%) received amikacin, 11/50 (22%) received capreomycin and 10/50 (20%) received streptomycin or a combination; reflecting a difference in policy between centres. Only 21/50 (42%) patients received baseline screening by audiogram within 2 weeks of starting treatment and 16/50 (32%) then had monthly audiograms, with the majority screened more infrequently and 12/50 (24%) receiving no screening. (ii) Of the 50 patients, 14 (28%) experienced ototoxicity, with 9/50 (18%) left with long-term hearing loss. Increased age (P¼0.02), use of amikacin (P¼0.02) and decreased renal function (P¼0.01) were significantly associated with ototoxicity. Conclusions: There is local variation in both the choice of injectable agent and in ototoxicity screening practices. Long-term morbidity from injectable treatment is significant even in this well-resourced setting, and the data suggest capreomycin might be associated with less ototoxicity when compared with amikacin. There is a need for more high-quality clinical data to inform future guidelines for treatment and monitoring. Keywords: aminoglycosides, capreomycin, drug toxicity Introduction The World Health Organisation (WHO) estimates that 5.3% of all tuberculosis cases are now multidrug-resistant tuberculosis (MDR-TB) and that these numbers are likely to increase in the coming years. 1 Patterns of international migration mean that MDR-TB will be an increasing challenge in both resource-rich and -poor settings. 2,3 Current international guidelines recommend a treatment course of at least 18 months with a combination of agents. 4 The aminoglycoside family of antimicrobials or the mechanistically similar cyclic peptide antibiotic capreomycin are recommended to form part of all MDR-TB treatment regimens. 4,5 Although a new drug for MDR-TB treatment, TMC207, is well advanced in clinical development, 6 injectable antimicrobials are likely to remain a key part of treatment for the foreseeable future. There is little high-quality randomized evidence to inform guidelines. WHO guidelines advise the use of amikacin or kanamycin initially, with capreomycin in resistant cases, and streptomycin is to be avoided due to widespread resistance, 4 although opinions differ. 5 All four agents are injectable, with the suggested duration of treatment being at least 6 months. 4 Therefore # The Author Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please journals.permissions@oup.com 1815

2 Sturdy et al. MDR-TB patients often receive a very large cumulative dose of injectable agents with significant toxicity. Ototoxicity and nephrotoxicity are both well-recognized adverse effects of injectable treatment. Ototoxicity can present with either hearing disturbance or vestibular symptoms. Hearing loss is classically irreversible, bilateral and high frequency, progressing to the lower frequencies in some patients, and is often accompanied by tinnitus. 7,8 Both patient and drug factors influence ototoxicity, 9 and in recent years there has been interest in the role of host genetics predisposing to ototoxicity, with the discovery that the A1555G mitochondrial DNA mutation is associated with some cases of aminoglycoside-induced hearing loss. 10,11 To better inform guidelines, we carried out the first multicentre survey of practice focused on injectable MDR-TB treatment and ototoxicity screening practices in the UK, with a further analysis of the incidence of drug-induced toxicity and the factors associated with its development. Methods A retrospective study was carried out at five UK treatment centres, four within greater London (St Mary s Hospital, Paddington; Northwick Park Hospital, Harrow; Royal Free Hospital, Hampstead; St George s Hospital, Tooting) and Heartlands Hospital, Birmingham. Centres were included if they had treated more than five cases in the period of study. Eligible cases were suspected or culture-proven MDR-TB patients treated between 1 January 2004 and 31 December 2009 and receiving an injectable agent for.2 weeks. Across the five centres, records were unavailable for eight eligible individuals. For each patient, the following baseline demographics were collected: age at the start of treatment, sex, ethnic group, HIV status, co-morbidities and any other concomitant medications, and the choice of injectable, administration route, duration of treatment, total dose and the reason for stopping injectable treatment were identified. Total dose was calculated from the dose, the frequency and the duration of treatment. Duration was defined as the number of days from initiation of injectable treatment until its withdrawal. For patients who received more than one injectable during treatment, the duration and dose are the combined total for both compounds. Laboratory data for baseline serum creatinine were recorded. Consistent with other studies, nephrotoxicity was then defined as an increase from baseline of 44 mmol/l 9,12 at any time during treatment. Data were collected from routine clinical records and anonymized, thus ethical approval was not required. All audiograms were collected and analysed. Audiograms were recorded using standard British Society of Audiology guidelines. 13 Most were performed in soundproof test conditions, although a small number were performed at the bedside. Patient-reported hearing loss or tinnitus at any point was recorded with the date first reported and whether symptoms persisted after treatment. Outcome variables were defined as cases with evidence of ototoxicity and cases without. Two definitions of ototoxicity were used: an audiogram-based definition and a clinical definition. Ototoxicity on audiogram was defined as a 20 db loss from baseline at any one test frequency or a 10 db loss at any two adjacent test frequencies. 14 An audiogram was defined as a true baseline result if it was taken within 2 weeks of the commencement of treatment. If there was no baseline, the first audiogram taken during treatment was used as the baseline. If a patient had only one audiogram, or had audiograms taken only after treatment had finished, ototoxicity was defined as two or more values,20 db. 12 The clinical definition used was cases with hearing loss or tinnitus persisting after treatment, without available audiograms. Factors associated with ototoxicity were investigated using binary logistic regression. All statistical analysis was carried out using SPSS for Windows v17 (SPSS Inc., Chicago, IL, USA, 2009). Results Demographics Fifty patients treated with injectable antimicrobials for MDR-TB were included in the analysis. Seven (14%) were from site 1, 8 (16%) from site 2, 12 (24%) from site 3, 14 (28%) from site 4 and 9 (18%) from site 5. Baseline characteristics are shown in Table 1. Creatinine results were unavailable for five patients. One case had baseline creatinine.110 mmol/l. This patient was in pre-dialysis renal failure due to unrelated co-morbidities before beginning injectable treatment. The same patient was the only case sustaining a creatinine.44 mmol/l above baseline after treatment cessation, having shown further deterioration. Injectable agent choice Twenty-nine (58%) of the group received amikacin, 11 (22%) capreomycin and 4 (8%) streptomycin as the first choice of injectable. Of the remaining six patients, five received both amikacin and capreomycin, with two switching from amikacin to capreomycin (one due to ototoxicity and one after transferring hospitals), two from capreomycin to amikacin (one due to pancytopenia and one for unspecified reasons) and one receiving both simultaneously. One remaining patient received capreomycin and streptomycin simultaneously. The variability in choice of injectable was largely due to variation in practice at different Table 1. Baseline demographics (percentages in parentheses) Variable Value Mean age+sd (years) Sex female 20 (40) male 30 (60) Ethnic group Asian Indian/Pakistani 13 (26) Asian Chinese 2 (4) Asian Other 6 (12) Black African 17 (34) Caucasian 10 (20) all other 2 (4) HIV status positive 5 (10) negative 45 (90) Baseline creatinine mean+sd (mmol/l) Creatinine increase.44 mmol/l from baseline at any point yes 7 (14) no 38 (76) not known 5 (10) 1816

3 MDR-TB treatment in the UK JAC Table 2. Injectable use by MDR centre (% within each centre in parentheses) Table 3. Audiogram screening by MDR centres (% within each centre in parentheses) MDR Centre AMK CAP STR AMK/CAP CAP/STR Total 1 2 (29) 4 (57) 1 (14) (100) (75) 3 (25) (14) 7 (50) 1 (7) 3 (22) 1 (7) (89) 1 (11) 9 Total AMK, amikacin; CAP, capreomycin; STR, streptomycin. centres, with two sites using capreomycin preferentially and three using amikacin (Table 2). Patients received injectables for a mean duration of 118 days (SD 82.8) with a dose of 92.2 g (SD 60.3). Dosing regimens varied, with patients receiving doses of between 750 and 1000 mg either once daily or three times weekly, following British National Formulary (BNF) guidelines of 15 mg/kg/day. 15 All centres monitored amikacin levels routinely in all patients, although frequency of testing and whether peak or trough measurements were taken varied between centres. A number of patients (17/50, 34%) had their initial dosing regimen changed at least once over the course of treatment due to drug levels, toxicity concerns or practical difficulties. Twenty-seven (55%) received treatment intravenously (iv) throughout (by slow infusion over 30 or 60 min), 13 (27%) intramuscularly (im) and 8 (16%) received by both administration routes over the course of treatment, with 1 patient receiving treatment by both iv and nebulizer and 1 patient s administration route was unknown. All streptomycin was administered im, with amikacin and capreomycin administered by both routes. Thirty-one (70.5%) patients stopped treatment after completing the agreed course, 10 (22.7%) stopped due to ototoxicity, 2 transferred abroad, and 1 died, with details for 6 patients unknown. Screening Thirty-eight patients (76%) received at least one audiogram, with screening practice varying between centres. Using the definition of baseline as an audiogram taken within 2 weeks of the commencement of treatment, 21/50 (42%) patients had baseline screening. Sixteen patients (32%) received monthly follow-up audiograms, with the majority screened less frequently or only after the report of symptoms. Twelve patients (24%) received no audiogram screening at any point. Table 3 shows these results by centre. Centres 1 and 4 performed the fewest number of audiograms per patient and were also the centres that preferentially used capreomycin. No centre routinely screens for the A1555G mutation or other mitochondrial mutations. Ototoxicity outcomes Fourteen patients (28%) showed evidence of ototoxicity. In 12/14 patients this was based on audiological definitions, while 2/14 reported persistent hearing disturbance with no available MDR Centre Patients Number receiving 1 AG Number receiving BL within 2 weeks Number receiving monthly AG Mean AGs received/ patient (29) 0 (0) 0 (0) (100) 5 (63) 6 (75) (83) 4 (33) 3 (25) (71) 5 (36) 2 (14) (89) 7 (78) 5 (56) 3.8 Total (76) 21 (42) 16 (32) AG, audiogram; BL, baseline. Loss from baseline at 8 khz R ear (db) Case 4 Case 9 Case Time from start treatment (months) Figure 1. Pattern of loss in three patients with a series of at least four audiograms recorded monthly. Dotted line indicates point at which patient first reported symptoms and at which treatment was stopped. audiograms. None of these patients is recorded as having reported symptoms before treatment began. Of the 12 cases with audiogram evidence of ototoxicity, 11 showed bilateral and 1 unilateral loss, with all cases showing loss at high frequency (.4 khz) and 6 with additional loss at speech frequencies ( khz). In total, 9/14 (64%) of the ototoxicity cases (18% of the total group) were left with symptomatic loss after treatment: 6/9 with tinnitus, 1/9 with hearing loss and 2/9 with both. Two patients were later known to have been given hearing aids. Of the 5/14 patients classified as suffering from ototoxicity but without persistent symptoms, 2/5 had audiogram changes that met our criteria with transient symptoms and 3/5 had audiogram changes with no reported symptoms. The pattern of audiogram change was further examined in three patients with hearing loss and a series of at least four audiograms recorded monthly (Figure 1). Loss of.20 db from baseline at high frequency (8 khz) was not seen in any of these three patients until at least 3 months after treatment began. Factors associated with ototoxicity Table 4 shows the outcome of the analysis of variables for the ototoxicity group compared with the non-ototoxicity group. 1817

4 Sturdy et al. Table 4. Univariate analysis of variables associated with ototoxicity in the total cohort (percentages shown in parentheses; P,0.05 taken as significant) Ototoxicity (%) n¼14 Table 5. Outcome by injectable used (% by outcome in parentheses) Outcome AMK CAP STR AMK/CAP CAP/STR Total Ototoxicity 11 (78.6) 0 1 (7.1) 2 (14.3) 0 14 No ototoxicity 18 (50) 11 (30.6) 3 (8.3) 3 (8.3) 1 (2.8) 36 Total AMK, amikacin; CAP, capreomycin; STR, streptomycin. No ototoxicity (%) n¼36 P value univariate Mean age+sd (years) Gender male 10 (71.4) 20 (55.6) female 4 (28.6) 16 (44.6) 0.3 Ethnicity Caucasian 4 (28.6) 6 (16.7) Asian Indian/Pakistani 2 (14.3) 11 (30.6) Asian Chinese 0 (0) 2 (5.6) Asian Other 1 (7.1) 5 (13.9) Black African 6 (42.9) 11 (30.6) all other 1 (7.1) 1 (2.8) 0.57 HIV status positive 1 (7.1) 4 (11.1) negative 3 (92.9) 32 (88.9) 0.68 Amikacin at any point yes 13 (92.9) 21 (58.3) no 1 (7.1) 15 (41.7) 0.02 Capreomycin at any point yes 2 (14.3) 15 (41.7) no 12 (85.7) 21 (58.3) 0.07 Duration of injectable treatment, mean+sd (days) Total dose of injectable agent, mean+sd (g) Baseline creatinine, mean (mmol/l) Creatinine increase.44 mmol/l from baseline at any point yes 5 (35.7) 2 (6.5) no 9 (64.3) 29 (93.5) not known Increased age, creatinine.44 mmol/l at any point and amikacin use were significantly associated with ototoxicity on analysis. Table 5 shows a further analysis of results according to choice of aminoglycoside. Discussion We present data from five UK centres as part of the first multicentre survey of current practice focused on monitoring the use of injectable antimicrobials for MDR-TB treatment. In total, 28% (14/50) of our patients suffered ototoxicity. These findings are within the range of four other MDR-TB ototoxicity studies that used audiogram-based definitions and found incidences ranging from 18% to 42%. 9,12,16,17 The variation in choice of injectable between the five centres was unexpected, with three centres showing a clear preference for amikacin and two for capreomycin. This may reflect some of the contradictions within current guidance a recent review argued for the preferential use of capreomycin (to avoid the development of aminoglycoside cross-resistance), 5 while the WHO guidelines recommend initial use of amikacin or kanamycin. 4 Several in vitro studies have compared the efficacy of capreomycin and amikacin, consistently concluding that although both are strongly bactericidal, amikacin is slightly more active than capreomycin, although the relevance of minor differences in in vitro MIC and in vivo activity is unclear. Guidance regarding audiological screening within the MDR-TB population also appears limited. Only the American Thoracic Society (ATS) has issued specific guidelines, and these recommend a baseline audiogram with further screening if symptoms are reported. 21 Practice in the centres studied differed from this, with only 42% of patients receiving a baseline audiogram, but two centres screened the majority of patients on a monthly basis. Current guidelines do not indicate how changes in a subject s audiogram should be managed. This is important because if treatment is not altered until a patient is symptomatic, the damage is often irreversible. 7,8 We found evidence of one patient who was showing significant high-frequency audiogram changes 2 months before she became symptomatic (Figure 1). Importantly, this suggests that a window of opportunity to modify treatment may exist in some patients after audiogram changes are apparent, but before symptoms appear and that this is not always currently being acted upon. This study does have notable limitations. We were unable to include eight eligible cases across the five centres. Although this represents a small proportion of individuals (14%), it could lead to some bias in results. There is no suggestion that these individuals had a worse outcome, and if all were asymptomatic, the overall proportion of patients with ototoxicity would fall to 14/58 (24.1%). It is possible that the centres included in this study are not representative of national practice as a whole, however, a comprehensive survey of practice was beyond the scope of this work. Also, audiogram coverage was not universal, meaning that for some cases the reliance was solely on these patients having reported symptoms and these having been subsequently recorded in the notes. Particularly of note, the two centres that preferentially used capreomycin performed the lowest number of audiograms (0.7 per patient at centre 1 and 1.7 at centre 4), which may have biased the capreomycin/amikacin comparison. However, records were extensive in all cases, and all centres followed up patients regularly and asked specifically about hearing disturbance, suggesting that bias in terms of clinical definition should be minimal. Despite this, there are a number of implications from this work. The scale of use of capreomycin and the associated 1818

5 MDR-TB treatment in the UK JAC decreased incidence of ototoxicity was unexpected. Given that the only comparison of efficacy between these two compounds is from in vitro studies, and that there are no head-to-head trials, we would advocate a randomization within future clinical trials to strengthen the evidence base and inform guidelines regarding which injectable agent should be used preferentially. This study was not able to assess a difference in clinical outcomes or the incidence of other serious toxicities that have been associated with capreomycin, such as hypokalaemia. 22 Capreomycin remains expensive for programmes in high-prevalence areas, and if better evidence supported its use, it would be possible to make a stronger advocacy argument to reduce drug prices. Data here from centres where routine audiograms are performed suggest they might help prevent symptomatic ototoxicity, which can be permanent. This will require both regular audiogram screening and an agreed, evidence-based threshold at which to review treatment. Although more data are needed to provide stronger conclusions, the evidence here suggests the need for baseline screening, with monthly screening at least from month 3 for those at highest risk of ototoxicity (e.g. the elderly and those with renal impairment) and a review of treatment if there is a loss of more than 20 db at any test frequency. Although additional measures, including routine testing for the A1555G mutation, may play an important role, other simpler measures are likely to have an even greater effect. In conclusion, MDR-TB is an increasingly common challenge in clinical practice and current injectable therapies can lead to substantial long-term ototoxicity. Better use of current treatment and monitoring strategies could reduce this. More detailed prospective studies and randomized data comparing agents within this class are needed to improve our knowledge base, potentially within the context of trials of new agents. Acknowledgements We thank Peter Davies for his critical appraisal of the manuscript. Funding This study was funded in part by the comprehensive Biomedical Research Centre (BRC) of Imperial College NHS Trust. Transparency declarations None to declare. References 1 World Health Organisation. Anti-tuberculosis drug resistance of the world. 2008; WHO/HTM/TB/ tions/2008/drs_report4_26feb08.pdf (1 November 2010, date last accessed). 2 O Riordan P, Schwab U, Logan S et al. Rapid molecular detection of rifampicin resistance facilitates early diagnosis and treatment of multi-drug resistant tuberculosis: case control study. PLoS One 2008; 3: e Abubakar I, Moore J, Drobniewski F et al. Extensively drug-resistant tuberculosis in the UK: 1995 to Thorax 2009; 64: World Health Organisation. Guidelines for the programmatic management of drug-resistant tuberculosis. 2008; WHO/HTM/TB/ eng.pdf (1 November 2010, date last accessed). 5 Caminero JA, Sotgiu G, Zumla A et al. Best drug treatment for multidrug-resistant and extensively drug-resistant tuberculosis. Lancet Infect Dis 2010; 10: Diacon AH, Pym A, Grobusch M et al. The diarylquinoline TMC207 for multidrug-resistant tuberculosis. N Engl J Med 2009; 360: Rizzi MD, Hirose K. Aminoglycoside ototoxicity. Curr Opin Otolaryngol Head Neck Surg 2007; 15: Guthrie OW. Aminoglycoside induced ototoxicity. Toxicology 2008; 249: Peloquin CA, Berning SE, Nitta AT et al. Aminoglycoside toxicity: daily versus thrice-weekly dosing for treatment of mycobacterial diseases. Clin Infect Dis 2004; 38: Viljoen DL, Sellars SL, Beighton P. Familial aggregation of streptomycin ototoxicity: autosomal dominant inheritance? J Med Genet 1983; 20: Prezant TR, Agapian JV, Bohlman MC et al. Mitochondrial ribosomal RNA mutation associated with both antibiotic-induced and non-syndromic deafness. Nat Genet 1993; 4: de Jager P, van Altena R. Hearing loss and nephrotoxicity in long-term aminoglycoside treatment in patients with tuberculosis. Int J Tuberc Lung Dis 2002; 6: British Society of Audiology. Recommended procedure: pure tone air and bone conduction threshold with and without masking and determination of uncomfortable loudness audiometry levels. 2004; (1 April 2011, date last accessed). 14 American Speech-Language-Hearing Association. Audiologic management of individuals receiving cochleotoxic drug therapy [guidelines]. 1994; (1 November 2010, date last accessed). 15 British National Formulary. BNF 58. London: BMJ Group and RPS Publishing, Duggal P, Sarkar M. Audiologic monitoring of multi-drug resistant tuberculosis patients on aminoglycoside treatment with long term follow-up. BMC Ear Nose Throat Disord 2007; 7: Torun T, Gungor G, Ozmen I et al. Side effects associated with the treatment of multidrug-resistant tuberculosis. Int J Tuberc Lung Dis 2005; 9: Heifets L, Lindholm-Levy P. Comparison of bactericidal activities of streptomycin, amikacin, kanamycin, and capreomycin against Mycobacterium avium and M. tuberculosis. Antimicrob Agents Chemother 1989; 33: Ho YI, Chan CY, Cheng AF. In-vitro activities of aminoglycoside-aminocyclitols against mycobacteria. J Antimicrob Chemother 1997; 40: Rastogi N, Labrousse V, Goh KS. In vitro activities of fourteen antimicrobial agents against drug susceptible and resistant clinical isolates of mycobacterium tuberculosis and comparative intracellular activities against the virulent H37Rv strain in human macrophages. Curr Microbiol 1996; 33: American Thoracic Society, CDC, and Infectious Diseases Society of America. Treatment of tuberculosis. MMWR Recomm Rep 2003; 52: Shin S, Furin J, Alcantara F et al. Hypokalemia among patients receiving treatment for multidrug-resistant tuberculosis. Chest 2004; 125:

6 Sturdy et al. 23 Bardien S, Human H, Harris T et al. A rapid method for detection of five known mutations associated with aminoglycoside-induced deafness. BMC Med Genet 2009; 10: Human H, Hagen CM, Jong GD et al. Investigation of mitochondrial sequence variants associated with aminoglycoside-induced ototoxicity in South African TB patients on aminoglycosides. Biochem Biophys Res Commun 2010; 393: Bardien S, de Jong G, Schaaf HS et al. Aminoglycosideinduced hearing loss: South Africans at risk. S Afr Med J 2009; 99: