{"_asset":{"@search.score":1,"Name":"Adherence, tolerability, and outcome after 36 months of isoniazid-preventive therapy in 2 rural clinics of Swaziland: a prospective observational feasibility study","AlternateTitleCommaDelimited":null,"Id":"5359","AssetLocationUri":null,"Abstract":"Although efficacy of 36 months isoniazid preventive therapy (IPT) among HIV-positive individuals has been proven in trial settings, outcome, tolerance, and adherence have rarely been evaluated in real-life settings. This is a prospective observational cohort study conducted in 2 primary care rural clinics in Swaziland. After negative tuberculosis symptom screening, patients either with the positive tuberculin skin test (TST) or after tuberculosis treatment were initiated on IPT for 144 weeks. In addition to routine clinic visits, adherence was assessed every semester. Of 288 eligible patients, 2 patients never started IPT (1 refusal, 1 contraindication), and 253 (87.8%), 234 (81.3%), and 228 (79.2%) were still on IPT after 48, 96, and 144 weeks, respectively (chiP = .01). Of 41 patients who interrupted IPT before 144 weeks, 21 defaulted (of which 17 also defaulted HIV care); 16 stopped because of adverse drug reactions; 2 were discontinued by clinicians' mistake and 1 because of TB symptoms. Five patients (1.7%) died of causes not related to IPT, 5 (1.7%) developed TB of which 2 were isoniazid-resistant, and 9 (3.1%) were transferred to another clinic. As an indicator of adherence, isoniazid could be detected in the urine during 86.3% (302/350) and 73.6% (248/337) of patient visits in the 2 clinics, respectively (chiP < .001). The routine implementation of IPT 36 months was feasible and good patient outcomes were achieved, with low TB incidence, good tolerance, and sustained adherence.","PublicationOrPresentationDate":"2017-09-01T00:00:00Z","LastModifiedDate":"2021-12-08T04:27:31.342Z","IsDeleted":false,"DigitalObjectIdentifier":"10.1097/MD.0000000000007740","ContentTypeName":"Research","ParentContentTypeName":"Journal Article","JournalName":"Medicine","ContentType":"Journal Article > Research","AvailabilityTypeId":1,"AvailabilityType":"Full Text","JournalIssue":"35","JournalVolume":"96","JournalPages":"e7740","JournalAbbreviation":"Medicine (Baltimore)","Conference":null,"ConferenceStartDate":null,"ConferenceEndDate":null,"content":"\nObservational Study Medicine®\n\nOPEN\nAdherence, tolerability, and outcome after\n36 months of isoniazid-preventive therapy\nin 2 rural clinics of Swaziland\nA prospective observational feasibility study\nYolanda Mueller, MD, MIHa,∗, Qhubekani Mpala, MScb, Bernhard Kerschberger, MD, MPHb,\nBarbara Rusch, MIHb, Gugu Mchunu, MPHc, Sikhathele Mazibuko, MDd, Maryline Bonnet, MD, PhDa,e\n\nAbstract\nAlthough efficacy of 36 months isoniazid preventive therapy (IPT) among HIV-positive individuals has been proven in trial settings,\noutcome, tolerance, and adherence have rarely been evaluated in real-life settings.\nThis is a prospective observational cohort study conducted in 2 primary care rural clinics in Swaziland.\nAfter negative tuberculosis symptom screening, patients either with the positive tuberculin skin test (TST) or after tuberculosis\n\ntreatment were initiated on IPT for 144 weeks. In addition to routine clinic visits, adherence was assessed every semester.\nOf 288 eligible patients, 2 patients never started IPT (1 refusal, 1 contraindication), and 253 (87.8%), 234 (81.3%), and 228 (79.2%)\n\nwere still on IPT after 48, 96, and 144 weeks, respectively (chi2P= .01). Of 41 patients who interrupted IPT before 144 weeks, 21\ndefaulted (of which 17 also defaulted HIV care); 16 stopped because of adverse drug reactions; 2 were discontinued by clinicians’\nmistake and 1 because of TB symptoms. Five patients (1.7%) died of causes not related to IPT, 5 (1.7%) developed TB of which 2\nwere isoniazid-resistant, and 9 (3.1%) were transferred to another clinic. As an indicator of adherence, isoniazid could be detected in\nthe urine during 86.3% (302/350) and 73.6% (248/337) of patient visits in the 2 clinics, respectively (chi2 P< .001).\nThe routine implementation of IPT 36months was feasible and good patient outcomeswere achieved, with low TB incidence, good\n\ntolerance, and sustained adherence.\n\nAbbreviations: ADR= adverse drug reaction, ART= antiretroviral therapy, HIV= human immunodeficiency virus, INH= isoniazid,\nIPT = isoniazid preventive therapy, IQR = interquartile range, PLHIV = people living with HIV, PY = person-years, TB = tuberculosis,\nTST = tuberculin skin test, WHO = World Health Organization.\n\nKeywords: adherence, isoniazid preventive therapy, TB-HIV\n[3]\n1. Introduction\n\nIn settings with high tuberculosis (TB) burden, isoniazid (INH)\nPreventive Therapy (IPT) is recommended for prevention of\nactive tuberculosis in people living with HIV (PLHIV),\nparticularly in individuals who are tuberculin skin test (TST)-\npositive.[1–3] In addition to antiretroviral therapy (ART), IPT\nreduces the incidence of active TB[4] and was initially\nEditor: Wenyu Lin.\n\nFunding: This study was funded by Médecins Sans Frontières (MSF). MSF was\ninvolved in all stages of the study conduct and analysis.\n\nThe authors have no conflicts of interest to disclose.\na Epicentre, Paris, France, bMédecins Sans Frontières, Geneva, Switzerland,\nc National Tuberculosis Control Program, d Swaziland National Aids Program,\nMinistry of Health, Swaziland, e IRD UMI 233 TransVIHMI-UM—INSERM U1175,\nMontpellier, France.\n∗\nCorrespondence: Yolanda Mueller, Department of Ambulatory Care and\n\nCommunity Medicine, University of Lausanne, Lausanne, Switzerland\n(e-mail: yolanda.mueller@hospvd.ch).\n\nCopyright © 2017 the Author(s). Published by Wolters Kluwer Health, Inc.\nThis is an open access article distributed under the Creative Commons\nAttribution License 4.0 (CCBY), which permits unrestricted use, distribution, and\nreproduction in any medium, provided the original work is properly cited.\n\nMedicine (2017) 96:35(e7740)\n\nReceived: 1 February 2017 / Received in final form: 26 June 2017 / Accepted:\n13 July 2017\n\nhttp://dx.doi.org/10.1097/MD.0000000000007740\n\n1\n\nrecommended for 6 months. However, effectiveness of IPT\nhas been shown to diminish 6 to 18 months after completion of\nsuch a regimen,[5–7] suggesting that IPT should be provided for\nlonger. Indeed, a trial conducted in Botswana showed that TB\nincidence among TST-positive PLHIV was reduced by a further\n43% if IPT was given for 36 months as compared to 6 months,[8]\n\nalthough the reduction was not significant in PLHIV with\nnegative or unknown TST results.\nDespite these results, uptake of 36 months’ IPT has been slow\n\ninmany countries.[9–11] Main concerns about long-term IPTwere\nrisk of side-effects, patient adherence, development of active TB\nwhile on IPT, and increased risk of INH-resistance. For example,\nin South Africa, up to 28% of cases suffered from a grade 3 or 4\ntransaminase elevation, after a median of one-and-a-half year on\nIPT and only 60% were retained on IPT at 3 years.[12] Besides,\nalthough benefits of IPT are greatest in TST-positive patients, the\nintroduction of widespread TST in resource-limited settings is\nlimited by operational constraints such as cold chain require-\nments and the extra-visit required for TST reading. Although\nclinical trials have been essential to inform recommendations\nabout 36 months IPT, there have been no observational studies\ndocumenting actual implementation. Confirming results from\nclinical trials in routine settings can help to convince decision-\nmakers about the added value of 36 months IPT.\nEven though TB incidence has decreased in recent years,\n\nSwaziland is still the country with the highest TB andHIV burden\nreported in the world.[13] National TB control has focused on\n\nmailto:yolanda.mueller@hospvd.ch\nhttp://creativecommons.org/licenses/by/4.0\nhttp://creativecommons.org/licenses/by/4.0\nhttp://dx.doi.org/10.1097/MD.0000000000007740\n\n\nMueller et al. Medicine (2017) 96:35 Medicine\nearly detection and treatment of TB, in parallel with rapid scaling\nup and integration of ART. Although IPT was recommended for\n6 months, its implementation remained challenging in 2011, and\nextending IPT duration to 36 months seemed unfeasible. There\nwas also the concern that adherence and outcomes under routine\nconditions would be less favorable than those reported in clinical\ntrials. Therefore, the objective of this study was to assess the\nfeasibility of TST-based 36 months IPT strategy in HIV-infected\npatients in programmatic conditions in 2 rural clinics of the\nShiselweni region in Swaziland, in order to inform policy-makers\nin the country and in similar settings. This paper focuses on\nadherence, tolerability, and outcomes after 48, 96, and 144\nweeks of follow-up. The results of this study on initial patient\nscreening, implementation of tuberculin-skin test, and initiation\nof IPT have been published previously.[14]\n2. Methods\n\nThis prospective observational cohort included PLHIV eligible\nfor IPT followed in 2 rural nurse-led primary care clinics in the\nShiselweni region, Kingdom of Swaziland (OLOS clinic in the\nHlatikulu zone and New Haven clinic in the Matsanjeni zone).\nThe study was undertaken in compliance with the Declaration of\nHelsinki and was approved by the Social Welfare Scientific\nEthical Committee of Swaziland and the Médecins Sans\nFrontières Ethical Review Board. Written informed consent\nwas obtained from all study participants.\nDetails of study setting and screening process can be found\n\nelsewhere.[14] Briefly, both clinics served an active ART cohort of\nabout 700 patients. Nurses were in charge of providing HIV-TB\ncare, which was fully integrated into general outpatient care.\nExpert clients (PLHIV trained to provide patient support\neducation and counseling) assisted the nurses in managing\npatient files and clinic registers, as well as adherence support.\nMedical doctors visited the clinic once per week or every 2 weeks,\naccording to needs, to attend to complicated cases and offer\ntechnical support.\nThe sample size was calculated to estimate the proportion of\n\npositive TST.[14] The IPT study population, derived from the\nproportion of positive TST, consisted of all at-least-16-years-old\nPLHIV enrolled for HIV care at 1 of the 2 study clinics who were\neligible for 36months’ IPT based on either a positive TST (≥5mm\ninduration) or for secondary prophylaxis (after TB treatment[2]).\nExclusion criteria were kept to a minimum to be as close as\npossible to the routine program conditions. According to\nnational recommendations for IPT provision, patients did not\nreceive IPT when they were on TB treatment, where TB was\nsuspected, or when INH was clinically contraindicated (e.g.\nhepatitis and allergy). Informed consent was obtained before TST\nscreening.[14]\n\nIntensified TB case-finding was implemented at every contact\nbetween a PLHIV and the health facility, based on symptom\nscreening (cough, night sweats, fever, weight loss, or chest pain)\nas per national recommendations.[15,16] Diagnosis of TB was\ninitially based on auramine-stained microscopical examination\nand XpertMTB/RIF (Cepheid, Sun Valley) testing performed at\nhealth-center level (secondary level), and later only on\nXpertMTB/RIF as per national recommendations.[17] Samples\nfrom smear- or Xpert-positive patients were sent to reference\nlaboratories at regional or national level for mycobacterial\nculture and drug susceptibility testing.\nPatients were seen after 1 month after starting IPT.\n\nSubsequently, appointments followed the usual visit schedule\n2\n\nfor ART or cotrimoxazole refill, that is, every 1, 2, or 3 months\ndepending on patient, clinic, and drug stocks. Laboratory\nmonitoring followed the general recommendations for patients\nunder HIV care in Swaziland, namely 6-monthly CD4, and\nquarterly liver function test. In 2012, viral-load testing was added\nto routine monitoring as per the national VL algorithm. Expert\nclients traced patients on HIV care by phone when clinical\nappointments were missed by 3 days. A renewed attempt was\nmade by phone in cases where patients did not present at the\nclinic over a period of 90 days. IPT outcome definitions were\nadapted from the national standard definitions used to monitor 6\nmonths’ IPT:[16] “default” was defined as a patient who had\ntaken INH for at least 1 month, then interrupted for 60\nconsecutive days or more; “failure” was defined as someone\ndeveloping active TB disease while on IPT; “treatment\ndiscontinuation” was used for someone whose IPT had been\ndiscontinued by a health care worker due to adverse effects or any\nother reason; “transfer out” was used for someone who\ntransferred to another ART site or region to continue treatment;\n“death”were those on IPTwhowere reported to have died of any\ncause during the course of treatment; “completed IPT” was\ndefined as a full course of 36 times 4 weeks (total 144 weeks) of\nINH.\nPill count was systematically recorded at every patient contact.\n\nIn addition, for all patients on IPT for 36 months, an in-depth\nevaluation was conducted twice yearly, consisting of a urine\nsample and a short questionnaire. Adherence was assessed\nthrough a combination of 4 parameters: pill count (i.e., the\nnumber of INH tablets remaining in the INH blister when coming\nfor drug refill, including refill for antiretroviral or other\nconcomitant drug), the self-reported number of days of no\nINH intake in the past 14 days, the Morisky scale, and isoniazid\ntesting in the urine. For this, isoniazid strip tests (BD TAXO INH\nTest) were performed on urine samples following the manu-\nfacturer’s instructions. This test is mostly reliable for the presence\nof isoniazid in the urine if the drug was taken in the last 24\nhours.[18] The Morisky scale[19] combines 4 questions related to\ndifferent aspects of drug intake by patients on chronic treatment\n(1. Do you ever forget to take your medicine? 2. Are you careless\nat times about taking your medicine? 3. When you feel better do\nyou sometimes stop taking your medicine? 4. Sometimes if you\nfeel worse when you take the medicine, do you stop taking it?).\nOne point is attributed for every positive answer. Adherence is\njudged high, intermediate, or low for scores of 0, 1 to 2, and 3 to\n4, respectively.[19]\n\nData were collected by the clinic staff on regular patient files\nduring the routine follow-up visits, before extraction on study-\nspecific case report forms by a study nurse. Occurrence of adverse\ndrug reactions (ADR) was recorded in the patient file by the clinic\nnurse. Adverse drug reactions were defined as: “a response to a\nmedicine which is noxious and unintended, and which occurs at\ndoses normally used in man.” Additional information was\nrequested in case of ADR, death, active TB, or treatment\ndiscontinuation. In addition, during adherence counseling, expert\nclients asked about side-effects experienced by the patients,\nwithout further precision on the definition of side-effect.\nData were entered on site weekly using Epi-Data 3.0 software\n\n(The EpiData Association, Odense Denmark), initially by a data\nentry clerk and later by the study nurse. Data were analyzed in\nStata 12.1 software (College Station, TX). No imputation was\ndone in the case of missing variables, apart from the date of death\nwhich was chosen randomly between date of last visit and date of\ninformation about deaths, in the absence of other information.\n\n\n\nMueller et al. Medicine (2017) 96:35 www.md-journal.com\nMain outcome was described by the proportions that completed\ntreatment, failed (developed TB), died, defaulted, discontinued,\nor transferred. The nonparametric test for trend across ordered\ngroups was used to assess changes on adherence indicators over\ntime. In the survival analysis of retention on IPT, death, TB,\ndefault and discontinuation were failures and transfers were\ncensored. Predefined baseline characteristics that could represent\nrisk factors for discontinuation (potential confounding factors),\nsuch as clinic, age, sex, ART status, initial baseline CD4 count at\nIPT initiation, and occurrence of adverse events, were tested in a\nnon-parametric survival model, and rates of interruption per\nperson-years were calculated. Variables associated with retention\non IPT, defined as a log rank test <0.20, were further included in\na multivariate Cox model in addition to age as a potential\nconfounder.\n\n3. Results\n\n3.1. Patient characteristics\n\nThe study profile has been previously published.[14] Patient\ncharacteristics are detailed in Table 1. Overall 288 patients\ninitiated IPT for the following reasons: 217 tested positive at the\nfirst TST, 47 tested positive at the second repeated TST at 6\nmonths after first TST, and 24 had completed TB treatment less\nthan 6 months ago. Median age was 38 years (interquartile range\n[IQR] 31–47). The median time from HIV diagnosis to inclusion\nwas 2.8 years (IQR: 1.4–4.8), with 252 patients (87.5%) already\non ART. Median CD4 count at inclusion was 477cell/microL\n(IQR 344 – 616). A total of 102 (35.9%) had been treated for\nTable 1\n\nCharacteristics of patients started on isoniazid preventive therapy, S\n\nOLOS\nN=152\n\nIPT indication\nFirst TST positive 116 (76.3)\nSecond TST positive 19 (12.5)\nSecondary prophylaxis 17 (11.2)\n\nCharacteristics at inclusion\nMedian age, IQR 39.5 (32–48)\nGender\nMale 47 (30.9)\nFemale 105 (69.1)\n\nOn ART, n, % 142 (93.4)\nMedian duration on ART N=141\nin years, IQR 2.1 (0.9–3.6)\n\nMedian time since HIV diagnosis N=141\nin years, IQR 2.7 (1.4–4.5)\n\nInitial CD4, at HIV diagnosis N=145\nMedian, IQR 214 (119–312)\n\nCD4 at inclusion N=152\nMedian, IQR 417 (322–579)\n\nCharacteristics at screening visit\nNumber of females N=105\nn pregnant, % 3 (2.9)\nWeight, kg N=143\nMean, SD 65.7 (13.2)\nTB and IPT history N=148\nPast TB treatment 66 (44.6)\nPast IPT 6 mo 72 (49.0)\n\nART = antiretroviral therapy, HIV = human immunodeficiency virus, INH = isoniazid, IPT = isoniazid p\n\n3\n\ntuberculosis previously, and 109 (38.5%) had previously\nreceived IPT for 6 months.\n3.2. Outcomes\n\nOut of a total of 288, 286 (99.3%) actually initiated IPT (1\nrefusal, 1 contraindication not previously recognized) and 253\n(87.8%), 234 (81.3%), and 228 (79.2%) were still on IPT after\n48, 96, and 144 weeks, respectively (chi2 P= .01, Table 2). Of the\n40 patients who did not complete IPT, 21 defaulted (of which 17\nalso defaulted HIV care), 16 discontinued IPT because of adverse\ndrug reactions (detailed below), and 2 because of a mistake of the\nclinician, who stopped IPT after 6 months instead of 36 months.\nNine patients (3.1%) were transferred to another clinic, and\n5 (1.7%) started TB treatment (described in detail below).\nFinally, 5 patients (1.7%) died while on IPT, corresponding to a\ndeath rate was 0.7/100 person-years (5 deaths/702 person-years\nof follow-up; 95%CI 0.3–1.7%). Of those 5 deaths, there was\n1 road traffic accident, 2 deaths of unknown cause, and 2 possibly\ndue to TB (1 patient with symptoms compatible with TB\nmeningitis, and 1 with important weight loss); none was\njudged possibly or probably related to IPT by the physician in\ncharge.\nFive patients were diagnosed with TB over 702 person-years of\n\nfollow-up, corresponding to an annual incidence rate of 0.7/100\nperson-years (95%CI 0.3–1.7/100PY). Among them, there were\n2 new pulmonary drug-sensitive cases, 1 extrapulmonary case\n(TB of the spine diagnosed by radiography), and 2 patients\ninfected with INH-resistant mycobacteria.\nwaziland, 2012–2015.\n\nNew Haven Both clinics\nN=136 N=288 P\n\n.048\n101 (74.3) 217 (75.4)\n28 (20.6) 47 (16.3)\n7 (5.2) 24 (8.3)\n\n37 (30–46) 38 (31–47) .192\n.694\n\n45 (33.1) 92 (31.9)\n91 (66.9) 196 (68.1)\n110 (80.9) 252 (87.5) .001\n\nN=108 N=249\n2.4 (1.1–4.5) 2.2 (1.0–3.8) .052\nN=127 N=268\n\n2.8 (1.4–5.0) 2.8 (1.4–4.8) .455\nN=130 N=275\n\n245 (133–371) 223 (127–334) .056\nN=135 N=287\n\n532 (390–678) 477 (344–616) <.001\n\nN=89 N=196\n9 (2.2) 5 (2.6) .770\nN=136 N=279\n\n64.6 (11.7) 65.2 (12.5) .466\nN=136 N=184\n\n36 (26.5) 102 (35.9) .001\n37 (27.2) 109 (38.5) <.001\n\nreventive therapy, IQR = interquartile range, TB = tuberculosis, TST = tuberculin skin test.\n\nhttp://www.md-journal.com\n\n\nTable 2\n\nCumulative outcomes after 48, 96, and 144 weeks on isoniazid preventive therapy, Swaziland, 2012–2015.\n\nOLOS (N=152) New Haven (N=136) Both clinics (N=288)\n\nW48 W96 W144 W48 W96 W144 W48 W96 W144\n\nOutcome\nIPT ongoing/completed 128 (84.2) 118 (77.6) 114 (75.0) 125 (91.9) 116 (85.3) 114 (83.8) 253 (87.8) 234 (81.3) 228 (79.2)\nDefault\n\n∗\n6 (4.0) 9 (5.9) 11 (7.2) 3 (2.2) 9 (6.6) 10 (7.4) 9 (3.2) 17 (5.9) 21 (7.3)\n\nFailure, active TB† 1 (0.7) 2 (1.3) 2 (1.3) 1 (0.7) 3 (2.2) 3 (2.2) 2 (0.7) 5 (1.7) 5 (1.7)\nDeath 3 (2.0) 3 (2.0) 3 (2.0) 1 (0.7) 1 (0.7) 2 (1.5) 4 (1.4) 4 (1.4) 5 (1.7)\nDiscontinuation\n\n∗\n11 (7.2) 15 (9.9) 16 (10.5) 4 (2.9) 4 (2.9) 4 (2.9) 15 (5.2) 19 (6.6) 20 (6.9)\n\nTransfer out 3 (2.0) 5 (3.2) 6 (4.0) 2 (1.5) 3 (2.2) 3 (2.2) 5 (1.7) 8 (2.8) 9 (3.1)\n∗\nIncluding 1 patient who never started IPT.\n\n† One patient treated for TB during the first year of follow-up finally diagnosed with a mycobacteria other than M tuberculosis. Final outcome was classified as discontinuation.\nIPT = isoniazid preventive therapy, TB = tuberculosis.\n\nMueller et al. Medicine (2017) 96:35 Medicine\nPatients were followed up for a median of 149 weeks (IQR\n146–163 weeks) after IPT start, for a total of 702 person-years\n(Fig. 1). In the adjusted non-parametric analysis (Table 3), gender\nandART treatmentwere not associatedwith retentionon IPT,while\nage, a highCD4count at IPT initiation, absence of adverse event and\nclinic were associated with better retention. In a multivariate Cox\nmodel, occurrence of at least 1 adverse event was the strongest\npredictor of treatment interruption, whereas increasing age and\nhigher initial CD4 level were associated with retention under IPT\n(P< .001 for goodness-of-fit test). The proportional hazards\nassumption was not violated (global test P= .325).\n\n3.3. Tolerance\n\nAdverse drug reactions were mentioned during consultations by\n5.6% (15/286), 2.0% (5/253), and 0.9% (1/234) patients during\nthe first, second, and third year of follow-up, respectively. Three\ncases of hepatotoxicity were recorded (1.0%), of which 2 were\nsevere (1 with ALT 5–10x, 1 admitted for jaundice) and 1 mild.\nAll patients recovered. Other patients mentioned symptoms\nwhich may have been due to hepatotoxicity, such as abdominal\npain (3 patients), or nausea and vomiting (2 patients, including 1\npregnant women). Unfortunately, IPT was often stopped without\nliver tests to confirm the diagnosis. The other reported adverse\nreactions were 2 cases of rash, 1 of itching, 2 cases of neuropathy,\n3 cases of nonspecific weight loss, 1 psychotic decompensation,\nand 1 case of vertigo and nystagmus.\n0 48 96 144\nWeeks on IPT\n\nclinic = 1 clinic = 2\n\nFigure 1. Kaplan–Meier curve of retention under IPT, by clinic.\n\n4\n\nTreatment was discontinued for 16/22 patients reporting\nadverse-drug reactions. Of the 16 patients that discontinued\nIPT, 1 was found dead at home 2 months after IPT\ndiscontinuation, and 1 pregnancy resulted in stillbirth. The 2\npatients with neuropathy recovered only partially. The remain-\ning 13 patients were seen at a later appointment with no\ncomplaints raised.\nSide-effects reported to the expert client during the adherence\n\nassessments decreased from 19.0% (16/79) and 50.0% (25/52) at\n6 months in OLOS andNewHaven, respectively, to less than 1%\nafter 30 months. These side-effects were mostly itching,\nsymptoms of neuropathy (tingling in hands or feet), nonspecific\npain (joint pains), and other complaints (abdominal pain). None\nled to temporary treatment interruptions. Although the nurses\nreported few cases of neuropathy (1.4%), up to 4.9% (14/288)\nwere reported during the interviews with the expert clients.\n\n3.4. Adherence\n\nAdherence was assessed at least once in 252 patients (87.5%),\nwith a median of 3 assessments per patient (IQR 2–4). Urine was\ncollected between 1 and 30hours after reported INH intake\n(median 13 hours in OLOS and 5 hours in New Haven, ranksum\ntest P< .001). Not all urine samples were actually tested, mainly\nbecause of the lack of available urine strips or of the staff either\ncollecting the urine or performing the test. Of all assessments\ncombined, the proportion of positive urine tests for isoniazid\n(INH) was 80.1% (95%CI 76.9–82.9%). Of the 17 patients who\nhad the urine tested more than 24hours after last INH intake, 14\n(82.4%) were still positive. By contrast, out of the 141 negative\nurine tests, there were only 11 patients admitting no INH intake\nin the past 24hours.\nAlthough objective assessments such as the proportion of\n\npositive urine tests or pill count showed decreasing adherence\nover time (test for trend P= .04 and .12, respectively, Table 4),\ninterview-based adherence indicators, such as the Morisky scale,\nor reported days of no intake seemed to improve over time. The\nmain reason given for missed drug intake was not having the drug\nwith them, or forgetting to take it.More rarely, patients cited lack\nof food, not feeling well, and running out of drugs before the\nappointment date.\n\n4. Discussion\n\nTo our knowledge, this is the first study to document feasibility of\n36 months’ IPT in the real life conditions of a high HIV/TB burden\nsetting. In this setting, patient outcomes were similar or even more\n0.00 0.25 0.50 0.75 1.00\n\n\n\nTable 4\n\nAdherence to isoniazid preventive therapy at M6, M12, M18, M24, M30, and M36, by pill count, intake in past 14 days, Morisky scale, and\nurine tests, Swaziland, 2012–2015.\n\nM6 M12 M18 M24 M30 M36\nN=131 N=121 N=166 N=147 N=84 N=128 Test for trend\n\nn % n % n % n % n % n % P\n\nUrine INH test\nNumber collected 117 89.3 109 90.1 155 93.4 128 87.1 79 94.0 115 89.8\nPositive 96 85.0 87 80.6 124 80.5 104 83.2 58 75.3 81 74.3 .041\nNegative 17 15.0 21 19.4 30 19.5 21 16.8 19 24.7 28 25.7\n\nMedian pill count, IQR\n∗\n\n7 3 to 14 6 0 to 6 8 3 to 15 9 3 to 13 8.5 4 to 17 9 3 to 13 .116\nNo intake past 24 h 6 4.6 9 7.4 6 3.6 8 5.4 2 2.4 0 0.0 .017\nDays of no intake in past\n\n2 weeks, 1 missing\n.005\n\n0 days 102 77.9 87 71.9 137 82.5 122 83.0 69 82.1 113 88.3\n1 day 15 11.5 21 17.4 17 10.2 12 8.2 7 8.3 5 3.9\n2 days 7 5.3 3 2.5 5 3.0 8 5.4 2 2.4 5 3.9\n3 days or more 7 5.3 10 8.3 7 4.2 5 3.4 6 7.1 5 3.9\n\nMorisky scale\nDo you ever forget to take\n\nyour medicine?\n26 19.8 13 10.7 13 7.8 15 10.2 8 9.5 5 3.9\n\nAre you careless at times about\ntaking your medicine?\n\n30 22.9 26 21.5 20 12.0 8 5.4 3 3.6 0 0.0\n\nWhen feeling better, do you stop\ntaking your medicine?\n\n1 0.8 2 1.7 0 0.0 0 0.0 0 0.0 0 0.0\n\nIf you feel worse, do you stop taking it? 1 0.8 1 0.8 0 0.0 1 0.7 0 0.0 0 0.0\nAdherence based on the Morisky scale <.001\nLow 1 0.8 2 1.7 0 0.0 0 0.0 0 0.0 0 0.0\nIntermediate 41 31.3 31 25.6 28 16.9 21 14.3 8 9.5 5 3.9\nHigh 89 67.9 88 72.7 138 83.1 126 85.7 76 90.5 123 96.1\n\n∗\nThe number of INH tablets remaining in the INH blister when coming for drug refill, including refill for antiretroviral or other concomitant drug.\n\nINH = isoniazid, IQR = interquartile range.\n\nTable 3\n\nUnivariate and multivariate Cox analyses of predictors of isoniazid preventive therapy interruption, Swaziland, 2012–2015.\n\nUnivariate Multivariate\n\nn 100 PY Rate 95%CI HR 95%CI P aHR 95%CI P\n\nOLOS clinic 37 345 0.11 0.08–9.15 1\nNew Haven clinic 22 347 0.06 0.04–0.10 0.61 0.36–1.03 .061\nMale 22 223 0.10 0.06–0.15 1\nFemale 37 465 0.08 0.06–0.11 0.80 0.47–1.36 .424\nNot on ART at initiation 10 86 0.12 0.06–0.22 1\nOn ART at initiation 49 616 0.08 0.06–0.11 0.69 0.35–1.37 .313\nAge, per 10 years 0.98 0.96–1.00 .098 0.96 0.93–0.99 .002\nInitial CD4 count, per 100 0.85 0.74–0.96 .008 0.84 0.73–0.95 .008\nNo adverse event 41 663 0.06 0.05 – 0.08 1\nOccurrence of at least 1 adverse event 18 39 0.46 0.29 – 0.73 6.4 3.70–11.2 <.001 9.2 5.00–17.10 <.001\n\naHR = adjusted hazard ratio, ART = antiretroviral therapy, CI = confidence interval, HR = hazard ratio.\n\nMueller et al. Medicine (2017) 96:35 www.md-journal.com\nfavorable than results obtained in African clinical trials that\ncompared 36 months’ isoniazid with other IPT regimens.[8,12] In\nparticular, the incidence of TB in patients on IPT was comparable\nwithother studies, even thoughSwazilandcarries ahighTBburden.\nHowever, most patients on IPT were already on ART for some\ntime, had high CD4 counts, and therefore were at lower risk of TB\nthan ART-naive patients. Two TB cases were due to isoniazid-\nresistant mycobacteria, which is not surprising in a setting were\nwhere 13.2% of new cases and 45.2% of previously treated cases\nharbor INH resistance.[20] Interestingly, both age and a high initial\nCD4 count favored good retention on IPT, whereas the occurrence\nof adverse drug reactions led to treatment interruption.\n5\n\nOccurrence of adverse drug reactions, particularly hepatotox-\nicity, was comparable with findings from the Botswana trial.[21]\n\nBy contrast, there were more hepatotoxic events recorded in the\nclinical trial in South Africa[12] than found in our study, maybe\nrelated to a higher prevalence of underlying hepatitis B and\nalcohol use. Also, liver tests were performed systematically in\nclinical trials, in contrast to our setting where, despite existing\nrecommendations, systematic laboratory testing is rare. In our\nstudy, hepatotoxicity was more frequent during the first year on\nINH. In these cases, extending IPT from 6 to 36 months did not\nmuch increase the risk of hepatotoxicity. However, we cannot\nexclude substantial underreporting of adverse events in our\n\nhttp://www.md-journal.com\n\n\nMueller et al. Medicine (2017) 96:35 Medicine\nstudy, considering that almost all adverse events reported led to\ntreatment discontinuation. This may imply that some side-effects\nremained unreported.\nIndeed, few cases of neuropathywere reported by the nurses given\n\nthe number of reported instances of tingling, burning sensation, or\npain in the hands and feet to the expert clients. Moreover,\nmanagement of neuropathy was far from ideal, INH being often\nstopped instead of increasing vitamin B6. Training on these aspects\nshould not be neglected during implementation of INH.\nAdherence to INH was good, based on the results of the urine\n\ntest, even if it slightly decreasing over time. However, some\nselectionbias in adherence estimates is possible, as the patientswho\nmissed their appointment, and therefore urine testing, may also be\nthe least adherent. Nonadherent patients were probably also more\nlikely to have the drug discontinued because of vague side-effects,\nwhereas the adherent patients were continuing on treatment. Still,\nour findings show that patients can be maintained on INH in\nroutine conditionswith good adherence. Themain reason given by\npatients formissing a dosewas not having the drugwith them.This\nsuggests that aspects of drugmanagementwhileon travel shouldbe\nreinforced during counseling. Also, differences in packaging\nbetween IPT and ART make dispensing a difficult task for health\ncenter staff. Coformulation with cotrimoxazole would certainly\nsimplify IPT prescription.[22]\n\nIn the context of an observational cohort conducted in routine\nconditions, our study suffers from some limitations. First, we\ncannot exclude some selection in the study population because of\nnonexhaustive screeningand/orTST refusal. Second, the studywas\nconducted in only 2 clinics, which may not be fully representative\nof all the clinics in the country. Also, most patients were already\nART-experienced, and results may not be generalized to contexts\nwith many ART-naive patients. Third, we did not aim to compare\noutcomes of 36months to 6months or non-IPT provision, because\nthe efficacy of IPT has been established by other studies.[8,12] Also,\nthe benefits and harms of short- and long-term IPT have been\ndescribed previously and therefore were not evaluated.[12,21,22]\n\nFinally, the quality of the information obtained for TB cases or\nadverse reactions was limited by the fact that many patients were\nnot assessed, or onlypartly, bymedical doctors.On the other hand,\n1 strength was the limited number of missing data, considering the\nsimple study set-up.\nThis feasibility study which was conducted in the public sector\n\nin a resource constrained setting shows that implementation of 36\nmonths’ IPT under routine conditions leads to similar outcomes\nas have been demonstrated previously in clinical trials. Indeed,\nIPT given for 36 months was well tolerated, with a limited\nincidence of TB cases. IPT can easily be integrated into routine\nfollow-up of HIV-positive patients. Furthermore, adherence can\nbe maintained in the long run. These findings support the 2015\nupdatedWHO recommendation of giving at least 36 months IPT\nto adults and adolescents living with HIV, who have an unknown\nor positive tuberculin skin test (TST) status, in resource-\nconstrained settings with high TB incidence and transmission.\n\nAcknowledgments\n\nThe authors would like to thank the patients who participated in\nthis study, the staff, and expert clients working in the clinics, and\nthe entire MSF teams working in Swaziland. Special acknowl-\nedgment to Mathieu Bastard for statistical advice, to Clothilde\nRambaud-Althaus for support in quality monitoring, and to\n6\n\nBéatrice Vasquez and CalorineMekiedje for help in data cleaning\nand project supervision.\nReferences\n\n[1] World Health OrganizationInterim Policy on Collaborative TB/HIV\nActivities. WHO, Geneva, Switzerland:2004.\n\n[2] World Health OrganizationGuidelines for Intensified Tuberculosis Case-\nFinding and Isoniazid Preventive Therapy for People Living with HIV in\nResource-Constrained Settings. WHO, Geneva, Switzerland:2011.\n\n[3] World Health Organization, UNAIDSPolicy Statement on Preventive\nTherapy Against Tuberculosis in People Living with HIV. WHO,\nGeneva, Switzerland:1998.\n\n[4] Group TAS, Danel C, Moh R, et al. A trial of early antiretrovirals and\nisoniazid preventive therapy in Africa. N Engl J Med 2015;373:808–22.\n\n[5] Johnson JL, Okwera A, Hom DL, et al. Duration of efficacy of treatment\nof latent tuberculosis infection in HIV-infected adults. AIDS 2001;15:\n2137–47.\n\n[6] Quigley MA, Mwinga A, Hosp M, et al. Long-term effect of preventive\ntherapy for tuberculosis in a cohort of HIV-infected Zambian adults.\nAIDS 2001;15:215–22.\n\n[7] Den Boon S, Matteelli A, Ford N, et al. Continuous isoniazid for the\ntreatment of latent tuberculosis infection in people living with HIV. AIDS\n2016;30:797–801.\n\n[8] Samandari T, Agizew TB, Nyirenda S, et al. 6-month versus 36-month\nisoniazid preventive treatment for tuberculosis in adults with HIV\ninfection in Botswana: a randomised, double-blind, placebo-controlled\ntrial. Lancet 2011;377:1588–98.\n\n[9] Charles MK, Lindegren ML, Wester CW, et al. Implementation of\ntuberculosis intensive case finding, isoniazid preventive therapy, and\ninfection control (“three I’s”) and HIV-tuberculosis service integration in\nlower income countries. PloS One 2016;11:e0153243.\n\n[10] Getahun H, Granich R, Sculier D, et al. Implementation of isoniazid\npreventive therapy for people living with HIV worldwide: barriers and\nsolutions. AIDS 2010;24(suppl 5):S57–65.\n\n[11] Gupta S, Granich R, Date A, et al. Review of policy and status of\nimplementation of collaborative HIV-TB activities in 23 high-burden\ncountries. Int J Tuberc Lung Dis 2014;18:1149–58.\n\n[12] Martinson NA, Barnes GL, Moulton LH, et al. New regimens to prevent\ntuberculosis in adults with HIV infection. N Engl J Med 2011;365:\n11–20.\n\n[13] Haumba S, Dlamini T, Calnan M, et al. Declining tuberculosis\nnotification trend associated with strengthened TB and expanded HIV\ncare in Swaziland. Public Health Action 2015;5:103–5.\n\n[14] Huerga H, Mueller Y, Ferlazzo G, et al. Implementation and operational\nresearch: feasibility of using tuberculin skin test screening for initiation of 36-\nmonth isoniazid preventive therapy in HIV-infected patients in resource-\nconstrained settings. J Acquir Immune Defic Syndr 2016;71:e89–95.\n\n[15] Getahun H, Kittikraisak W, Heilig CM, et al. Development of a\nstandardized screening rule for tuberculosis in people living with HIV in\nresource-constrained settings: individual participant data meta-analysis\nof observational studies. PLoS Med 2011;8:e1000391.\n\n[16] Ministry of Health, Kingdom of Swaziland. National Guidelines for\nImplementing the “Three I’s” of Tuberculosis Control. 2012.\n\n[17] Sikhondze W, Dlamini T, Khumalo D, et al. Countrywide roll-out of\nXpert\n\n®\n\nMTB/RIF in Swaziland: the first three years of implementation.\nPublic Health Action 2015;5:140–6.\n\n[18] Elizaga J, Friedland JS. Monitoring compliance with antituberculous\ntreatment by detection of isoniazid in urine. Lancet 1997;350:1225–6.\n\n[19] Morisky DE, Green LW, Levine DM. Concurrent and predictive validity\nof a self-reported measure of medication adherence. Med Care 1986;24:\n67–74.\n\n[20] Sanchez-Padilla E, Dlamini T, Ascorra A, et al. High prevalence of\nmultidrug-resistant tuberculosis, Swaziland, 2009–2010. Emerg Infect\nDis 2012;18:29–37.\n\n[21] Tedla Z, Nyirenda S, Peeler C, et al. Isoniazid-associated hepatitis\nand antiretroviral drugs during tuberculosis prophylaxis in HIV-\ninfected adults in Botswana. Am J Respir Crit Care Med 2010;182:\n278–85.\n\n[22] Harries AD, Lawn SD, Suthar AB, et al. Benefits of combined preventive\ntherapy with co-trimoxazole and isoniazid in adults living with HIV: time\nto consider a fixed-dose, single tablet coformulation. Lancet Infect Dis\n2015;15:1492–6.\n\n\n\tAdherence, tolerability, and outcome after 36 months of isoniazid-preventive therapy in 2 rural clinics of Swaziland\n\t1 Introduction\n\t2 Methods\n\t3 Results\n\t3.1 Patient characteristics\n\t3.2 Outcomes\n\t3.3 Tolerance\n\t3.4 Adherence\n\n\t4 Discussion\n\tAcknowledgments\n\tReferences\n\n\n","UrlSlug":"adherence-tolerability-outcome-after-36-months-isoniazid-preventive-therapy-2-rural-clinics-swaziland-prospective-observational-feasibility-study","collections":[],"countries":[{"Name":"Eswatini","Id":71}],"asset_languages":[{"AlternateTitle":null,"language":[{"Id":1,"Name":"English"}],"contentTranslation":[],"contentLocation":[]}],"creators":[{"Name":"Bonnet MMB","Id":15195,"ac":[{"Order":6}]},{"Name":"Mchunu G","Id":17566,"ac":[{"Order":4}]},{"Name":"Mpala Q","Id":17567,"ac":[{"Order":1}]},{"Name":"Rusch B","Id":18074,"ac":[{"Order":3}]},{"Name":"Kerschberger B","Id":19962,"ac":[{"Order":2}]},{"Name":"Mazibuko S","Id":20805,"ac":[{"Order":5}]},{"Name":"Mueller YK","Id":28236,"ac":[{"Order":0}]}],"keywords":[{"Name":"tuberculosis","Id":3},{"Name":"models of care","Id":9}]}}

Science Portal

Adherence, tolerability, and outcome after 36 months of isoniazid-preventive therapy in 2 rural clinics of Swaziland: a prospective observational feasibility study