Abstract
Background
The introduction of antiretroviral therapy (ART) has resulted in significant mortality reduction and improvement in the quality of life. However, this has come at a cost of increased drug toxicity. The objective of this study was to assess the patterns and predictors of ART toxicity in adult HIV patients in Ethiopia.
Methods
This is a prospective cohort study conducted at seven teaching hospitals between September 2009 and December 2013 involving 3921 HIV patients on ART. Adverse drug reactions (ADR) due to ART were identified based on clinical assessment and/or laboratory parameters. Multivariable random effects Poisson regression analysis was used to identify factors independently associated with toxicity.
Result
ADR due to ART drugs was reported in 867 (22.1 %) of the participants; 374 (9.5%) had severe forms. About 87% of reported toxicities were limited to three organ systems — the skin, nervous system and blood. The overall incidence of ADR was 9 per 100 person years. About a third of toxicities occurred during the first six months after ART initiation with the incidence rate of 22.4 per 100 person years. Concomitant anti-tuberculosis treatment was the strongest independent predictor of toxicity.
Conclusion
ADR was found to be highly prevalent in HIV patients on ART at tertiary hospitals in Ethiopia. Most of these conditions occurred early after ART initiation and in those with concomitant anti-tuberculosis treatment. Thus, routine monitoring of patients on ART should be strengthened with particular emphasis in the first 6 months. Strategies should also be devised to replace older and more toxic agents with newer and safer drugs available.
Keywords: HIV, ART, adverse drug reaction, incidence rate, ACM, Ethiopia
Introduction
Human immunodeficiency Virus (HIV) has caused significant socioeconomic crisis for the last few decades in sub-Saharan Africa (1). The discovery of antiretroviral therapy (ART) was the breakthrough in care and treatment of people living with HIV, resulting in reduction in mortality and improvement in the quality of life. This has transformed HIV from untreatable condition into a chronic manageable disease (2). However use of ART and other drugs used in HIV care has resulted in an increment in serious adverse drug reactions (ADR) some of which are life-threatening conditions (3, 4).
Adverse drug reaction is a response to a drug which is noxious and unintended. It is drug toxicity which occurs at doses normally used for prophylaxis, diagnosis, or treatment (5). Among HIV patients, the possible causes for drug toxicity are antiretroviral (ARV) drugs themselves; drug used for prophylaxis and treatment of opportunistic infections (OI) and other concomitant medications (6, 7). Adverse reactions occur with all ARV drugs and are among the major reasons for switching or discontinuing therapy as well as for medication non-adherence (8–10). ADRs have thus become one of the most important limiting factors to the successful treatment of HIV (11–13).
The magnitude of the problem significantly differs between developed and developing world due to various reasons (7). Toxicities related to ART in the developing world, in sub-Saharan Africa in particular, are more prevalent because the regimens contain older and more toxic agents like stavudin (d4t), zidovudin (AZT) and nevirapin (NVP). In addition to these, patients late presentation with advanced disease state, occurrence of opportunistic infections and comorbid medical conditions like malnutrition have made the problem even worse (6, 7, 14). Lack of laboratory monitoring and skilled professionals may also result in failure or delay in diagnosing specific toxicities and hence increasing their severity (6, 7, 12).
In the year 2013, about 790,000 people were estimated to live with HIV in Ethiopia; with adult prevalence of 1.2% and 45,000 deaths annually (15). Routine HIV care and free antiretroviral therapy (ART) services started in the country in 2005 (16). This has resulted in a marked decrement in morbidities and mortalities related to the disease and improvement in quality of life. There are currently about 320,000 adults receiving ART in Ethiopia (17). Few studies in the country have shown that ARV drug toxicity in Ethiopia is a common occurrence and major reason for ART regiment change or modification (9, 18, 19). However, most of the studies lack generalizability due to small sample size focusing on limited facility in central Ethiopia. The aim of this study was thus to assess the magnitude and patterns of ARV drug toxicity in adult HIV patients at major tertiary hospitals in Ethiopia.
Methods
Study design and setting: This study is part of Advanced Clinical Monitoring (ACM) multi-center cohort. The ACM cohort was conducted in Ethiopia at seven teaching Hospitals located at various parts of the country (Addis Ababa, Hawassa, Jimma, Haramaya, Mekelle, Gondar and Army forces hospital). The cohort included both ART naïve and ART experienced adult and pediatric HIV positive patients.
This article presents findings of adverse drug reaction in adult HIV patients on ART. To involve all patients who were started on ART since the beginning of HIV care in Ethiopia, both retrospective and prospective cohort study designs were used. In the retrospective design, data was obtained from medical records of ART experienced patients since initiation of ART which covers the period of January 2005 to August 2009. The other group consisted of ART naïve patients who were enrolled since 2009. ART experienced patients who were still alive and on follow-up at the beginning of the prospective design were followed in the same way as prospective group.
Baseline data collected from participants included socio-demographic profile, general medical conditions, laboratory parameters and components of HIV care given to them (ART regimen, treatment for OI and prophylaxis given). Participants were subsequently followed for occurrence of adverse reactions related to ART using clinical parameters and laboratory data where applicable. A particular emphasis was given to the first 6 months after ART initiation as most of the toxicities tend to occur during this time (7). These data were obtained from medical records for patients who were started on ART before 2009 (before recruitment started).
Assessment for ADR was done by physician working at ART clinics at respective hospitals. Patients were assessed on each visit using checklists prepared for routine follow-up evaluation. National guidelines (and WHO recommendations if applicable) for timing of hematologic and blood chemistry evaluation were followed. In patients with suspected ADR, all necessary laboratory tests were done as needed.
Inclusion and Exclusion Criteria: To assess ART related adverse drug reactions, this study included all adult ACM cohort participants who were either ART naïve or ART experienced HIV positive patients at the start of the cohort. Only participants of age 18 years or older at enrollment were included in this study. Accordingly, a total of 3921 adult HIV positive patients from seven teaching hospitals in Ethiopia were included in this study for analysis of ART drug toxicity. Large proportion of the participants, 2304 (58.8%) belonged to ART experienced group.
Patients who died or were transferred out within a week of ART initiation and those who were on ART for the same duration at the last recruitment were not included in this study.
Definition of outcome variables: The outcome of interest in this study was drug toxicities. An algorithm was developed to identify drug toxicities meeting the operational definition based on World Health Organization (WHO) criteria for ADR (5) (Annex 1).
Adverse drug reactions in patients on ART were identified based on clinical and/or laboratory data. Only clinical conditions known to arise from potential agent, in the absence of other compelling medical conditions, were attributable to ADR. Mild medical conditions which commonly occur in the general public but can be attributable to the drugs were excluded from this list unless specified and there was clear evidence for direct contribution by the agent. In most cases this refers to gastrointestinal disturbances (dyspeptic symptoms) which in most of the patients have been reported before or after ART initiation. Major gastrointestinal upset reported by patient to be worsened after these medication or has resulted in their discontinuation due to these conditions were considered as evident drug toxicity.
Severe or life threatening toxicity was defined based on clinical or laboratory evidences classified as grade 3 or 4 by WHO criteria (Annex — 1). All forms of toxicities resulting in hospitalization, major disability, changing or discontinuation of the culprit agent were categorized in this group.
In cases of multiple occurrences of toxicities in a single patient, the count was based on the systems involved and time of occurrence. Multisystem complications that were due to single mechanism were counted as one (e.g lactic acidosis and pancreatitis due to stavudin were considered as related toxicities if occurred within one week of each other). However, if the pathophysiologic mechanisms were different, they were counted separately (e.g., anemia and lactic acidosis in zidovudin were considered as separate forms of toxicities). Similar toxicities, toxicities of similar pathophysiology or of the same spectrum occurring within one week of each other and non-resolving ones were counted as one type of toxicity.
All forms of toxicity were classified based on their severity and body system affected.
Data quality assurance: Data collectors were trained and regularly supervised by Johns Hopkins Technical Support for the Ethiopian HIV/AIDS Initiative (TSEHAI). Refresher trainings were given to sites based on gaps identified. Standard reporting formats were used to minimize interpersonal variation and Variation between sites. All incoming staffs (physicians and nurse) were trained on entry.
Data management and Data analysis: The data was initially computerized using Microsoft access. This data was exported to SPSS version 20 and STATA version 12 for analysis. Tableau version 8 was also used to portray the trends in toxicity.
Because toxicities are more likely to occur during the initial period of exposure to ART, we calculated the six-month incidence of drug toxicity as a probability of a participant experiencing one or more drug toxicities during the first six months from initiation of ART. The crude incidence rates of toxicity per 100 person-years on ART was calculated by summing the total number of occurrences of toxicity and dividing this total by the summed person-time of each participant from ART start to either the time patient left the study or study terminated.
Incidence of toxicity was estimated using random effects multivariable Poisson regression analysis. Potential risk factors to be included in multivariable model were selected if p-value in the bivariate analysis was less than 20%. Clustering at site level was adjusted both in univariate and multivariable analysis.
We also calculated stratified incidence rates for each year that patients took ART. It was acknowledge that each participant might have one or more incidence of toxicity. The six-month incidence and incidence rates of all toxicities combined and of each severity category was calculated separately. Homogeneity of the 6-month incidence across sites was assessed using the Breslow-Day test and homogeneity of the incidence rates across sites was assessed using a chi-square test.
Ethical considerations: Ethical clearance of the study was obtained from Ethiopian Public Health Institute, Minister of Science and Technology of Ethiopia and Center for Diseases Control and prevention (CDC), Atlanta, USA. Written informed consent was obtained from all participants before recruitment. Data was collected using numerical codes and will remain anonymous.
Results
Baseline characteristics: ACM cohort of 3921 adult HIV patients who were on ART were assessed for the occurrence of ART drug toxicity; 62% of them were women. The mean age at ART initiation was 34.5 years (SD =9.2) and 75% were between 25 and 44 years of age. More than 60% of the participants had baseline body weight of less than 55 kilograms. Nearly 60% of the participants reported use of some form of substance; 6.7% have used hard drugs including morphine and cocaine. The majority of the participants were at advanced staged of their HIV at initiation of ART; 65.9% had WHO stage ≥3 disease and 69.7% had CD4 count of <200. In terms of functional status, 4.4% of the patients were bedridden at baseline (Table 1).
Table 1.
Baseline characteristics of adult HIV patients at initiation of ART in HIV patients at tertiary hospitals in Ethiopia.
| Total participant | Toxicity | |||||
| Yes | No | |||||
| Age category (N=3921) | Number | % | Number | % | Number | % |
| <25 | 399 | 10.2 | 92 | 23.1 | 307 | 76.9 |
| 25–34 | 1761 | 44.9 | 386 | 21.9 | 1375 | 78.1 |
| 35–44 | 1181 | 30.1 | 248 | 21.0 | 933 | 79.0 |
| >=45 | 580 | 14.8 | 141 | 24.3 | 439 | 75.7 |
| Gender (N=3921) | ||||||
| Female | 2426 | 61.9 | 609 | 25.1 | 1817 | 74.9 |
| Male | 1495 | 38.1 | 258 | 17.3 | 1237 | 82.7 |
| Baseline body weight, Kg (N=3738) | ||||||
| <45 | 802 | 20.5 | 209 | 26.1 | 593 | 73.9 |
| 45–54 | 1547 | 39.5 | 359 | 23.2 | 1188 | 76.8 |
| 55–64 | 970 | 24.7 | 197 | 20.3 | 773 | 79.7 |
| ≥65 | 419 | 10.7 | 69 | 16.5 | 2350 | 83.5 |
| ART history (N=3921) | ||||||
| ART experienced | 2304 | 58.8 | 642 | 27.9 | 1662 | 72.1 |
| ART naive | 1617 | 41.2 | 225 | 13.9 | 1392 | 86.1 |
| Educational status (N=3843) | ||||||
| No education | 723 | 18.4 | 189 | 26.1 | 534 | 73.9 |
| Primary | 1393 | 35.5 | 304 | 21.8 | 1089 | 78.2 |
| Secondary | 1310 | 33.4 | 276 | 21.1 | 1034 | 78.9 |
| Tertiary | 417 | 10.6 | 78 | 18.7 | 339 | 81.3 |
| Employment status (N=3285) | ||||||
| Unemployed | 1571 | 40.1 | 374 | 23.8 | 1197 | 76.2 |
| Too sick to work | 385 | 9.8 | 92 | 23.9 | 293 | 76.1 |
| Working part-time | 91 | 2.3 | 20 | 22.0 | 71 | 78 |
| Working full time | 1225 | 31.2 | 230 | 18.8 | 995 | 81.2 |
| Other | 13 | 0.3 | 3 | 23.1 | 10 | 76.9 |
| Marital status (N=3870) | ||||||
| Never married | 663 | 16.9 | 122 | 18.4 | 541 | 81.6 |
| Married(Inc. de facto) | 1834 | 46.8 | 358 | 19.5 | 1476 | 80.5 |
| Separated or Divorced | 803 | 20.5 | 218 | 27.1 | 585 | 72.9 |
| Widow/Widower | 570 | 14.5 | 151 | 26.5 | 419 | 73.5 |
| Facility (Hospital) (N=3921) | ||||||
| Tikur Anbessa | 566 | 14.4 | 165 | 29.2 | 401 | 70.8 |
| Army | 557 | 14.2 | 53 | 9.5 | 504 | 90.5 |
| Gondar | 610 | 15.6 | 155 | 25.4 | 455 | 74.6 |
| Jimma | 540 | 13.8 | 83 | 15.4 | 457 | 84.6 |
| Mekelle | 582 | 14.8 | 178 | 30.6 | 404 | 69.4 |
| Haramaya | 530 | 13.5 | 129 | 24.3 | 401 | 75.7 |
| Hawassa | 536 | 13.7 | 104 | 19.4 | 432 | 80.6 |
| WHO staging (N=3852) | ||||||
| 1 and 2 | 1265 | 32.8 | 253 | 20.0 | 1012 | 80.0 |
| 3 and 4 | 2587 | 67.2 | 598 | 23.1 | 1989 | 76.9 |
| CD4 count category (N=3789) | ||||||
| <200 | 2735 | 72.2 | 581 | 21.2 | 2154 | 78.8 |
| 200–349 | 983 | 25.9 | 232 | 23.6 | 751 | 76.4 |
| ≥350 | 71 | 1.9 | 17 | 23.9 | 54 | 76.1 |
| Functional status (N=3840) | ||||||
| Working | 2796 | 71.3 | 631 | 22.6 | 2165 | 77.4 |
| Ambulatory | 872 | 22.2 | 199 | 22.8 | 673 | 77.2 |
| Bedridden | 172 | 4.4 | 23 | 13.4 | 149 | 86.6 |
| ART regiment at initiation (N=3918) | ||||||
| AZT+3TC+EFV | 434 | 11.1 | 87 | 20.0 | 347 | 80.0 |
| AZT+3TC+NVP | 1053 | 26.9 | 237 | 22.5 | 816 | 77.5 |
| D4T+3TC+EFV | 344 | 8.8 | 93 | 27.0 | 251 | 73.0 |
| D4T+3TC+NVP | 951 | 24.3 | 289 | 30.4 | 662 | 69.6 |
| TDF+3TC+EFV | 920 | 23.5 | 122 | 13.3 | 798 | 86.7 |
| TDF+3TC+NVP | 209 | 5.3 | 39 | 18.7 | 170 | 81.3 |
| Others | 7 | 0.18 | 0 | 0 | 7 | 100.0 |
The initial ART regimen in all patients was a combination of two nucleoside reverse transcriptase inhibitors (NRTI) and one non-nucleoside reverse transcriptase inhibitor (NNRTI). The most commonly prescribed NRTIs were AZT (38%) and d4t (33.1%). However, this trend changed after the introduction of tenofovir (TDF) in 2009 which was used as an initial NRTI in 53.6% of the cases since then. Efavirinez (EFV) and NVP were the only NNRTs prescribed; the latter was used in 56.5% of the patients. All but one patient took lamivudine (3TC) as the backbone of the entire regimen.
Over 90% (3585) of them were on cotrimoxazole prophylactic treatment (CPT) while only 4.1% (160) were prescribed isoniazid (INH) preventive therapy (IPT) at time of ART initiation; 15.4% (604) were taking anti-tuberculosis treatment.
The mean follow-up period of patients on ART was 43.7 (SD =28.7) months with median follow-up of 41.6 months. There was significant difference in the mean follow-up of patients in Experienced group (mean = 60.3; SD = 24.5; median = 64.8 months) and Naïve group (mean = 19.2; SD =13.1; median 18.5 months).
Patterns of adverse drug reaction: Overall, there were 1253 incidents of ADR reported during the follow-up period. Of these, 542 (43.3%) were severe or life threatening toxicities. The rate of toxicity was the highest in the first 6 months after ART initiation; 31.8% occurred during this period with the trend gradually decreasing overtime (Fig 1). In general, ADR was reported in 867 (22.1 %) of the participants over the follow-up period; 374 (9.5%) had severe forms. Of those with ADR, 240 (27.7%) had more than one incident.
Figure 1.
Trends in occurrence of ADR in adult HIV patients on ART at tertiary hospitals in Ethiopia
The incidence rate of ADR over follow-up time was 9 per 100 person years. The incidence of toxicities was highest in the first 6 months of ART initiation with incidence rate of 22.4 per 100 person years.
About 87% of reported toxicities were limited to three organ systems — the skin (36.6%), nervous system (35.4%) and blood (14.9%). While most of dermatologic ADR were milder forms, about 67.0% of severe forms of adverse events were neurologic (Fig 2). Moreover, over 80% of neurologic toxicities were classified as severe forms which have resulted in either disabilities or hospitalization or regimen change. Most of the patients with neurologic toxicities and all with severe forms were taking d4t at the time of the toxicity incidence.
Figure 2.
Types of ADR by system affected in HIV patients on ART at tertiary hospitals in Ethiopia
The incidence rate of adverse events in the first six months by year of ART start (2005 – 2013) ranged from 11.9 to 36.8 per 100 person year. In general, the 6 months and overall incidence rate of ADR were higher in patients who started on ART since the 2009 as compared those who started earlier (Table 2).
Table 2.
Incident Rate of Toxicity by Year of ART Start at six month and over all follow-up period in HIV patients at tertiary hospitals in Ethiopia.
| ART start year |
Within 6 month | Overall | |||||
| Follow Up Time (in days |
No. of Toxicity |
Incident rate (Per 100 Person-Year) |
Follow Up Time |
No. of Toxicity |
Incident rate (Per 100 Person-Year) |
||
| 2005 | 41378 | 21 | 18.5 | 586067 | 103 | 6.4 | |
| 2006 | 110812 | 36 | 11.9 | 1380500 | 277 | 7.3 | |
| 2007 | 92257 | 52 | 20.6 | 993457 | 256 | 9.4 | |
| 2008 | 79201 | 49 | 22.6 | 718597 | 174 | 8.8 | |
| 2009 | 85133 | 58 | 24.9 | 584458 | 168 | 10.5 | |
| 2010 | 114886 | 88 | 28.0 | 555865 | 152 | 10.0 | |
| 2011 | 58475 | 59 | 36.8 | 193012 | 84 | 15.9 | |
| 2012 | 52531 | 29 | 20.2 | 101332 | 33 | 11.9 | |
| 2013 | 14670 | 6 | 14.9 | 15669 | 6 | 14.0 | |
Predictors of adverse drug reaction: Multivariable random effect Poisson regression analysis was done to find predictors for occurrences of ADR in the first six months of HAART initiation. Patients who started ART in 2009 and after had higher rates of ADR as compared to those who started treatment prior to that; AIRR 1.38, 95% CI 1.017 – 1.87 and 1.49, 95% CI 1.03 – 2.15 for 2009–2010 and 2011–2013 respectively as compared to ART start year of 2005–2008. Similarly, lower literacy status, marital inconvenience (being divorced or separated), and concomitant anti-TB treatment were independent predictors of ADR in the first 6 months (Table 3).
Table 3.
Predictors of Toxicity within 6 months of ART initiation in HIV patients at tertiary hospitals in Ethiopia: both ART naïve and experienced groups.
| Characteristics | Uni-variable analysis | Multivariable analysis | |||
| IRR (95% CI) | P-value | IRR (95% CI) | P -Value | ||
| Sex | Female | 1 | 1 | ||
| Male | 0.803(0.640 – 1.007) | 0.057 | 0.825(0.630 – 1.079) | 0.160 | |
| Age at ART start |
18 –24 | 1 | |||
| 25 –34 | 1.159(0.795 – 1.688) | 0.443 | |||
| 35 –44 | 1.047(0.703 – 1.559) | 0.823 | NS | ||
| 45+ | 1.245(0.809 – 1.915) | 0.319 | |||
| ART Start Year |
2005–2008 | 1 | 1 | ||
| 2009–2010 | 1.524(1.202 – 1.932) | 0.000 | 1.378(1.017 – 1.868) | 0.039* | |
| 2011–2013 | 1.533(1.173 – 2.003) | 0.002 | 1.487(1.029 – 2.151) | 0.035* | |
| Educational status |
No education | 1 | 1 | ||
| Primary | 0.724(0.545 – 0.961) | 0.025 | 0.729(0.543 – 0.980) | 0.036* | |
| Secondary | 0.867(0.656 – 1.145) | 0.314 | 0.851(0.633 – 1.144) | 0.284 | |
| Tertiary | 0.807(0.545 – 1.194) | 0.283 | 0.866(0.573 – 1.309) | 0.495 | |
| Marital Status | Never married | 1 | 1 | ||
| Married(Inc. de facto) | 1.145(0.827 – 1.585) | 0.416 | 1.093(0.778 – 1.537) | 0.608 | |
| Separated or Divorced | 1.663(1.181 – 2.340) | 0.004 | 1.545(1.070 – 2.230) | 0.020* | |
| Widow/Widower | 1.027(0.687 – 1.534) | 0.898 | 1.100(0.719 – 1.683) | 0.659 | |
| Weight at ART Start |
<=45 | 1 | 1 | ||
| 46 –51 | 0.973(0.726 – 1.304) | 0.855 | 1.075(0.791 – 1.460) | 0.645 | |
| 52 –58 | 0.935(0.693 – 1.263) | 0.663 | 1.011(0.732 – 1.397) | 0.947 | |
| 59+ | 1.224(0.905 – 1.655) | 0.190 | 1.323(0.936 – 1.869) | 0.113 | |
| I | 1 | ||||
| Base WHO Stage |
II | 1.291(0.870 – 1.915) | 0.205 | ||
| III | 1.034(0.721 – 1.482) | 0.858 | NS | ||
| IV | 1.107(0.723 – 1.694) | 0.640 | |||
| Base Functional Status |
Working | 1 | 1 | ||
| Ambulatory | 0.935(0.715 – 1.221) | 0.621 | 1.019(0.759 – 1.370) | 0.899 | |
| Bedridden | 0.535(0.254 – 1.128) | 0.100 | 0.539(0.221 – 1.314) | 0.174 | |
| Base CD4 Count |
<=50 | 1 | |||
| 51 – 100 | 0.958(0.659 – 1.393) | 0.823 | |||
| 101 – 200 | 0.905(0.646 – 1.266) | 0.559 | NS | ||
| 201 – 350 | 1.240(0.879 – 1.750) | 0.220 | |||
| 350+ | 1.433(0.674 – 3.049) | 0.350 | |||
| Base Regimen |
TDF-NVP | 1 | 1 | ||
| TDF-EFV | 0.745(0.500 – 1.111) | 0.149 | 0.650(0.428 – 0.986) | 0.043* | |
| ZDV-EFV | 0.649(0.418 – 1.008) | 0.054 | 0.687(0.426 – 1.107) | 0.123 | |
| ZDV-NVP | 0.559(0.380 – 0.821) | 0.003 | 0.596(0.396 – 0.897) | 0.013* | |
| d4T-EFV | 0.443(0.260 – 0.756) | 0.003 | 0.551(0.304 – 0.997) | 0.049* | |
| d4T-NVP | 0.487(0.323 – 0.736) | 0.001 | 0.573(0.351 – 0.935) | 0.026* | |
| CPT | Yes | 0.864(0.602 – 1.239) | 0.426 | NS | |
| No | 1 | ||||
| IPT | Yes | 0.853(0.466 – 1.561) | 0.606 | NS | |
| No | 1 | ||||
| TB Treatment | Yes | 1.439(1.124 – 1.843) | 0.004 | 1.443(1.081 – 1.926) | 0.013* |
| No | 1 | 1 | |||
Statistically significant, NS – not statistically significant (P was > 0.25 in univariate analysis)
CPT – co-trimoxazole prophylactic therapy, IPT – Isoniazid prophylactic therapy, TB – tuberculosis
Similar analysis for patients started on HAART since 2009 (prospectively recruited patients) showed that lower educational status, lower CD4 count (<200), and concomitant TB treatment were independent predictors of ADR (Table 4).
Table 4.
Predictors of Toxicity within 6 months of ART initiation in HIV patients at tertiary hospitals in Ethiopia: ART naïve group.
| Characteristics | Uni-variable analysis | Multivariable analysis | |||
| IRR (95% CI) | P-Value | IRR (95% CI) | P-Value | ||
| Sex | Female | 1 | 1 | ||
| Male | 0.63 (0.446, 0.901) | 0.011 | 0.71 (0.465, 1.077) | 0.107 | |
| Age ART Start | 18 –24 | 1 | |||
| 25 –34 | 1.38 (0.744, 2.577) | 0.304 | |||
| 35 –44 | 1.08 (0.560, 2.063) | 0.828 | |||
| 45+ | 1.22 (0.601, 2.492) | 0.579 | |||
| ART Start Year | 2013 | 1 | 1 | ||
| 2009 | 1.60 (0.548, 4.680) | 0.389 | 1.20 (0.351, 4.126) | 0.768 | |
| 2010 | 1.66 (0.719, 3.819) | 0.235 | 1.51 (0.593, 3.849) | 0.387 | |
| 2011 | 2.12 (0.907, 4.973) | 0.083 | 2.15 (0.841, 5.511) | 0.110 | |
| 2012 | 1.30 (0.537, 3.136) | 0.563 | 0.96 (0.356, 2.591) | 0.937 | |
| Education | No education | 1 | 1 | ||
| Primary | 0.54(0.357, 0.839) | 0.006 | 0.52 (0.326, 0.822) | 0.005* | |
| Secondary | 0.83 (0.555, 1.233) | 0.351 | 0.83 (0.536, 1.290) | 0.410 | |
| Tertiary | 0.71 (0.416, 1.206) | 0.204 | 0.87 (0.500, 1.530) | 0.639 | |
| Marital Status | Never married | 1 | 1 | ||
| Married(Inc. de facto) | 1.32 (0.819, 2.130) | 0.254 | 1.34 (0.797, 2.251) | 0.270 | |
| Separated or Divorced | 1.72 (1.026, 2.881) | 0.040 | 1.75 (0.995, 3.066) | 0.052 | |
| Widow/Widower | 1.30 (0.704, 2.415) | 0.399 | 1.24 (0.628, 2.450) | 0.536 | |
| Weight at ART Start | <=45 | 1 | 1 | ||
| 46 –51 | 0.71 (0.452, 1.120) | 0.141 | 0.77 (0.472, 1.260) | 0.299 | |
| 52 –58 | 0.73 (0.471, 1.134) | 0.162 | 0.91 (0.558, 1.478) | 0.697 | |
| 59+ | 0.76 (0.486, 1.174) | 0.212 | 0.93 (0.557, 1.478) | 0.765 | |
| Base WHO Stage | I | 1 | 1 | ||
| II | 1.67 (1.004, 2.770) | 0.048 | 1.82 (1.056, 3.145) | 0.031* | |
| III | 1.48 (0.901, 2.417) | 0.122 | 1.46 (0.830, 2.559) | 0.189 | |
| IV | 1.68 (0.913, 3.109) | 0.095 | 1.62 (0.785, 3.346) | 0.191 | |
| Base Functional Status | Working | 1 | |||
| Ambulatory | 0.99 (0.617, 1.589) | 0.967 | |||
| Bedridden | 0.91 (0.288, 2.884) | 0.875 | |||
| Base CD4 Count | <=50 | 1 | 1 | ||
| 51 – 100 | 0.79 (0.473, 1.311) | 0.359 | 0.74 (0.425, 1.286) | 0.285 | |
| 101 – 200 | 0.62 (0.389, 1.000) | 0.050 | 0.57 (0.337, 0.972) | 0.039* | |
| 201 – 350 | 0.83 (0.520, 1.332) | 0.443 | 0.82 (0.474, 1.406) | 0.464 | |
| 350+ | 1.17 (0.474, 2.871) | 0.738 | 1.20 (0.450, 3.196) | 0.717 | |
| Base Regimen | TDF-NVP | 1 | 1 | ||
| TDF-EFV | 0.72 (0.448, 1.155) | 0.172 | 0.52 (0.318, 0.863) | 0.011* | |
| ZDV-EFV | 0.91 (0.502, 1.654) | 0.760 | 0.78 (0.415, 1.466) | 0.440 | |
| ZDV-NVP | 0.76 (0.458, 1.271) | 0.299 | 0.69 (0.403, 1.174) | 0.170 | |
| d4T-EFV | 0.74 (.169, 3.254) | 0.692 | 0.25 (0.033, 1.955) | 0.188 | |
| d4T-NVP | 0.83 (0.246, 2.790) | 0.762 | 0.79 (0.229, 2.759) | 0.719 | |
| CPT | Yes | 0.87 (0.503, 1.505) | 0.619 | ||
| No | 1 | ||||
| IPT | Yes | 0.48 (0.067, 3.413) | 0.461 | ||
| No | 1 | ||||
| TB Treatment | Yes | 1.72 (1.179, 2.499) | 0.005 | 1.67 (1.043, 2.677) | 0.033* |
| No | 1 | 1 | |||
Statistically significant
When severe forms of toxicity were analyzed separately, marital inconvenience (being divorced or separated) and concomitant TB treatment remained independent predictors. AZT-EVF containing regimen was also found to result in 71% decrement in severe toxicity as compared to TDF-NVP containing regimen (Table 5).
Table 5.
Predictors of Severe Adverse drug reactions within 6 months of ART initiation in HIV patients at tertiary hospitals in Ethiopia: ART experienced group.
| Characteristics | Uni-variable analysis | Multivariable analysis | |||
| IRR (95% CI) | P Value | AIRR (95% CI) | P Value | ||
| Sex | Female | 1 | NS | ||
| Male | 0.771(0.515 – 1.155) | 0.207 | |||
| Age ART Start |
18 –24 | 1 | 1 | ||
| 25 –34 | 1.125(0.550 – 2.303) | 0.747 | 1.142(0.556 – 2.345) | 0.717 | |
| 35 –44 | 1.011(0.475 – 2.154) | 0.977 | 0.977(0.453 – 2.105) | 0.953 | |
| 45+ | 2.147(1.013 – 4.550) | 0.046 | 2.099(0.971 – 4.539) | 0.059 | |
| ART Start Year |
2005–2008 | 1 | 1 | ||
| 2009–2010 | 0.663(0.429 – 1.026) | 0.065 | 0.785(0.454 – 1.356) | 0.385 | |
| 2011–2013 | 0.661(0.393 – 1.113) | 0.120 | 0.829(0.422 – 1.631) | 0.588 | |
| Educational status |
No education | 1 | 1 | ||
| Primary | 0.526(0.316 – 0.873) | 0.013 | 0.601(0.358 – 1.010) | 0.054 | |
| Secondary | 0.714(0.445 – 1.146) | 0.164 | 0.901(0.549 – 1.478) | 0.681 | |
| Tertiary | 0.739(0.386 – 1.417) | 0.363 | 0.975(0.501 – 1.897) | 0.941 | |
| Marital Status |
Never married | 1 | 1 | ||
| Married(Inc. de facto) | 1.057(0.584 – 1.913) | 0.855 | 1.073(0.587 – 1.962) | 0.818 | |
| Separated/Divorced | 2.107(1.157 – 3.838) | 0.015 | 2.110(1.141 – 3.901) | 0.017* | |
| Widow/Widower | 1.263(0.630 – 2.531) | 0.511 | 1.217(0.591 – 2.503) | 0.595 | |
| <=45 | 1 | ||||
| Weight at ART Start |
46 –51 | 0.984(0.584 – 1.658) | 0.952 | NS | |
| 52 –58 | 0.936(0.545 – 1.608) | 0.811 | |||
| 59+ | 1.120(0.649 – 1.932) | 0.684 | |||
| Base WHO Stage |
I | 1 | |||
| II | 0.992(0.455 – 2.160) | 0.984 | NS | ||
| III | 1.174(0.600 – 2.297) | 0.639 | |||
| IV | 1.432(0.673 – 3.048) | 0.351 | |||
| Base Functional Status |
Working | 1 | |||
| Ambulatory | 1.131(0.707 – 1.809) | 0.607 | NS | ||
| Bedridden | 0.762(0.240 – 2.418) | 0.645 | |||
| Base CD4 Count |
<=50 | 1 | |||
| 51 – 100 | 0.866(0.445 – 1.685) | 0.672 | |||
| 101 – 200 | 0.898(0.501 – 1.611) | 0.718 | |||
| 201 – 350 | 1.290(0.712 – 2.337) | 0.402 | |||
| 350+ | 0.472(0.062 – 3.559) | 0.466 | |||
| Base Regimen |
TDF-NVP | 1 | 1 | ||
| TDF-EFV | 0.769(0.336 – 1.759) | 0.534 | 0.544(0.232 – 1.277) | 0.162 | |
| ZDV-EFV | 0.493(0.190 – 1.281) | 0.147 | 0.285(0.104 – 0.779) | 0.014* | |
| ZDV-NVP | 0.741(0.340 – 1.615) | 0.450 | 0.622(0.277 – 1.396) | 0.249 | |
| d4T-EFV | 1.057(0.428 – 2.610) | 0.905 | 0.537(0.196 – 1.473) | 0.227 | |
| d4T-NVP | 1.217(0.557 – 2.656) | 0.623 | 0.817(0.335 – 1.988) | 0.655 | |
| CPT | Yes | 1.350(0.675 – 2.701) | 0.396 | ||
| No | 1 | ||||
| IPT | Yes | 1.619(0.705 – 3.719) | 0.256 | ||
| No | 1 | ||||
| TB Treatment |
Yes | 2.409(1.630 – 3.559) | 0.000 | 2.817(1.847 – 4.298) | <0.0001* |
| No | 1 | 1 | |||
Statistically significant
It is thus evident that anti-tuberculosis treatment in the first 6 months of HAART in initiation was found to be the strongest independent predictor of ADR; any form of ADR increased by 67% in those taking TB treatment (Table 3) while severe forms were found to be 3 times higher.
Discussion
Toxicities related to ART drug were widely prevalent among HIV patients on follow-up in tertiary hospitals in Ethiopia; significant proportion of them resulted in serious morbidities. Toxicities were found to be highest in the first 6 months of ART initiation. Concomitant anti-tuberculosis treatment was found to be an independent predictor of adverse drug reaction due to ART.
All of the participants were started on combination of two NRTIs and NNRTI adhering to WHO's recommendation for resource limited settings (20). Most of patients taking d4t were shifted to TDF since 2009 after revision of national guidelines in 2008 (16). As a result, the proportion of patients starting on d4t containing regimen dropped from 70% in 2006 to 1% in 2013. However, the fact that some patients were still started on stavudin containing regimen (42 patients in 2010, 4 in 2011, 3 in 2012 and 2 in 2013) years after revision of the guideline should not pass unnoticed.
The overall prevalence of ADR of 22.1% in our study is comparable with findings from the developing world; 19.5% in Cameroon (21), 26.8% in India (22). However, the incidence rate is much higher than findings in Nigeria of 4.6 per 100 person years (23). Because of the very nature of occurrences of ADR due to ART, neither assessment of the total prevalence of the problem nor the overall incidence rate tell the exact magnitude of the problem. Longer exposure to the drugs may inflate the prevalence and overall incidence rate of ADR. However, more serious and common toxicities occur during the initial phase of ART (7). Hence, comparison between studies may not be straight forward.
Consistent with global data, incident toxicities in our study were found to be more common in the first 6 months of ART initiation (4, 7, 24). Patients with advanced disease at ART initiation in particular are at higher risk of ADR (25, 26). However, neither WHO stage of the disease nor CD4 count at ART initiation was associated with occurrences of ADR in our study. This may be partly explained by the fact that most the study participants had advanced WHO stage and low CD4 count at HAART initiation. Additionally, failure to detect and report milder forms of ADR may explain the absence of this relationship. Despite using standardized reporting format for any ADR, it is well documented that both highly trained professionals and patient underreport milder forms (27, 28). We thus believe that milder dermatologic manifestations and gastrointestinal symptoms might have been underreported by patients and overlooked by treating physician as non-life threatening event.
Early toxicities are also known to be worsened with concomitant OI treatment (29). Though INH and Cotrimoxazole prophylaxis for OIs were not associated with increased toxicities in our study, concomitant anti-tuberculosis admiration was found to be the strongest predictor of ADR. Overlapping toxicities between HAART and anti-TB drugs has been well documented and know to hamper success of the treatment (30). While this has been an issue, the benefit of addressing the two infections should take priority.
When years of ART initiation were compared, incident toxicity was found to be more common in those who started ART since 2009. This difference is mainly due to design effect. Data for patients who started ART before 2009 was obtained only retrospectively and incident toxicities could have been under reported. When only prospectively cohorts were analyzed separately, the effect of year of ART initiation disappears. We believe that the lowest incidence of toxicities in prior years was mainly due to underreporting than other reasons. Besides, as experience in HIV care has improved over the last many years in the country, detection rate of all forms of adverse reaction has expectedly improved resulting in the increment in the rate of reported ADR.
Even though this study is multicenter large scale cohort in Ethiopia, there are certain limitations worth mentioning. First of all, the data prior to 2009 were collected from medical records only. As a result, certain ADR, the non-life threatening ones, could have been underreported. Secondly, some of the adverse reactions diagnosed on clinical ground only were based on treating physician's diagnosis which might have also resulted inconsistence across sites and under-reporting. Last but most importantly, the study was based in tertiary/university hospitals in the country with better laboratory facilities and more qualified professionals in HIV care. Hence, the findings may not be generalizable to all HIV patients in Ethiopia.
In conclusion, ADR due to ART drugs has been found to be highly prevalent in HIV patients on follow-up at tertiary hospitals in Ethiopia. A significant proportion was found to be serious adverse reactions resulting in admissions and/or regimen changes. ADRs were prominent during the first 6 months after ART initiation. Concomitant administration of full anti-tuberculosis regimen was the strongest independent predictor for the occurrence of these toxicities. Thus, routine clinical and laboratory monitoring of patients on ART should be strengthened with particular emphasis in the first 6 months. High risk patients (those with comorbidities, advanced immunosuppression and concomitant OIs) should be actively identified before ART initiation. These patients should be effectively treated for the OIs or comorbidities, if possible, before ART initiation. Selection of ART regimen should also put into consideration their other medical conditions. Strategies should also be devised to replace older and more toxic agents with newer and safer drugs available. All facilities rendering ART service should follow standard reporting method for incident ADR related to ART and OI treatment.
Acknowledgments
We are indebted to Johns Hopkins University, USA and United States Centers for Disease Control and Prevention (CDC) for funding this research and following its progress throughout. We also would like to thank the seven hospitals and their staffs for their support during the data collection. We also acknowledge EPHI for its leading role during manuscript preparation.
We would also like to express our acclaim to Dr Alula M. Teklu for his perseverance in taking this process to the level of publication.
References
- 1.Dixon S, McDonald S, Roberts J. The impact of HIV and AIDS on Africa's economic development. Bmj. 2002;324(7331):232–234. doi: 10.1136/bmj.324.7331.232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Mwagomba B, Zachariah R, Massaquoi M, Misindi D, Manzi M, Mandere BC, et al. Mortality reduction associated with HIV/AIDS care and antiretroviral treatment in rural Malawi: evidence from registers, coffin sales and funerals. PloS one. 2010;5(5):e10452. doi: 10.1371/journal.pone.0010452. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pau AK. Antiretroviral Therapy-associated Serious and Life-threatening Toxicities. Current infectious disease reports. 2003;5(5):429–438. doi: 10.1007/s11908-003-0023-9. [DOI] [PubMed] [Google Scholar]
- 4.Masenyetse LJ, Manda SO, Mwambi HG. An assessment of adverse drug reactions among HIV positive patients receiving antiretroviral treatment in South Africa. AIDS research and therapy. 2015;12:6. doi: 10.1186/s12981-015-0044-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.International drug monitoring: the role of national centres. Report of a WHO meeting. World Health Organization technical report series. 1972;498:1–25. [PubMed] [Google Scholar]
- 6.Murphy RA, Sunpath H, Kuritzkes DR, Venter F, Gandhi RT. Antiretroviral therapy-associated toxicities in the resource-poor world: the challenge of a limited formulary. The Journal of infectious diseases. 2007;196(Suppl 3):S449–S456. doi: 10.1086/521112. [DOI] [PubMed] [Google Scholar]
- 7.Subbaraman R, Chaguturu SK, Mayer KH, Flanigan TP, Kumarasamy N. Adverse effects of highly active antiretroviral therapy in developing countries. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America. 2007;45(8):1093–1101. doi: 10.1086/521150. [DOI] [PubMed] [Google Scholar]
- 8.O'Brien ME, Clark RA, Besch CL, Myers L, Kissinger P. Patterns and correlates of discontinuation of the initial HAART regimen in an urban outpatient cohort. Journal of acquired immune deficiency syndromes. 2003;34(4):407–414. doi: 10.1097/00126334-200312010-00008. [DOI] [PubMed] [Google Scholar]
- 9.Wube M, Tesfaye A, Hawaze S. Antiretroviral Therapy Regimen Change Among HIV/AIDS Patients in Nekemt Hospital: a Primary Care Hospital in Oromia Regional State, Ethiopia. Journal of Applied Pharmaceutical Science. 2013;3(8):5. [Google Scholar]
- 10.Njuguna C, Orrell C, Kaplan R, Bekker LG, Wood R, Lawn SD. Rates of switching antiretroviral drugs in a primary care service in South Africa before and after introduction of tenofovir. PloS one. 2013;8(5):e63596. doi: 10.1371/journal.pone.0063596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hartmann M. The side effects of antiretroviral therapy. Hautarzt. 2006;57(11):6. doi: 10.1007/s00105-006-1240-6. [DOI] [PubMed] [Google Scholar]
- 12.Yunihastuti E, Widhani A, Karjadi TH. Drug hypersensitivity in human immunodeficiency virus-infected patient: challenging diagnosis and management. Asia Pacific allergy. 2014;4(1):54–67. doi: 10.5415/apallergy.2014.4.1.54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Shet A, Antony J, Arumugam K, Kumar Dodderi S, Rodrigues R, DeCosta A. Influence of adverse drug reactions on treatment success: prospective cohort analysis of HIV-infected individuals initiating first-line antiretroviral therapy in India. PloS one. 2014;9(3):e91028. doi: 10.1371/journal.pone.0091028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mehta U, Durrheim DN, Blockman M, Kredo T, Gounden R, Barnes KI. Adverse drug reactions in adult medical inpatients in a South African hospital serving a community with a high HIV/AIDS prevalence: prospective observational study. British journal of clinical pharmacology. 2008;65(3):396–406. doi: 10.1111/j.1365-2125.2007.03034.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.UNAIDS, author. AIDS epidemic update: the latest epidemiology data. 2014. [cited 2014 30 August]. Available from: www.unaids.org. [Google Scholar]
- 16.Federal ministry of health, author. Guideline for the management of opportunistic infections and antiretroviral treatment in adolescents and adults in Ethiopia. Addis Ababa, Ethiopia: Federal HIV/AIDS Prevention and Control Office Federal Ministry of Health; 2008. [Google Scholar]
- 17.Global fund. Global fund mid-2014 results. 2014. [cited 2014 30 August]. Available from: http://www.theglobalfund.org/documents/results/Results_2014-07-17_Factsheet_en/ [Google Scholar]
- 18.Abdissa SG, Fekade D, Feleke Y, Seboxa T, Diro E. Adverse drug reactions associated with antiretroviral treatment among adult Ethiopian patients in a tertiary hospital. Ethiopian medical journal. 2012;50(2):107–113. [PubMed] [Google Scholar]
- 19.Jima Y, Angamo M, Wabe N. Causes for antiretroviral regimen change among HIV/AIDS patients in Addis Ababa, Ethiopia. Tanzania Journal of Health Research. 2013;15(1) doi: 10.4314/thrb.v15i1.2. [DOI] [PubMed] [Google Scholar]
- 20.Consolidated Guidelines on the Use of Antiretroviral Drugs for Treating and Preventing HIV Infection: Recommendations for a Public Health Approach. WHO Guidelines Approved by the Guidelines Review Committee. Geneva: 2013. [PubMed] [Google Scholar]
- 21.Namme Luma H, Doualla MS, Choukem SP, Temfack E, Ashuntantang G, Achu Joko H, et al. Adverse drug reactions of Highly Active Antiretroviral Therapy (HAART) in HIV infected patients at the General Hospital, Douala, Cameroon: a cross sectional study. The Pan African medical journal. 2012;12:87. [PMC free article] [PubMed] [Google Scholar]
- 22.Divakar B, Mistry S, Kantharia N, Mamtarani The Study of Adverse Drug Reactions (ADR's) in HIV patients taking Highly Active Antiretroviral Therapy in ART centre, NCH, Surat, India. Int J Med Pharm Sci. 2012;3(2):10. [Google Scholar]
- 23.Eluwa GI, Badru T, Agu KA, Akpoigbe KJ, Chabikuli O, Hamelmann C. Adverse drug reactions to antiretroviral therapy (ARVs): incidence, type and risk factors in Nigeria. BMC clinical pharmacology. 2012;12:7. doi: 10.1186/1472-6904-12-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Bezabhe WM, Bereznicki LR, Chalmers L, Gee P, Kassie DM, Bimirew MA, et al. Adverse Drug Reactions and Clinical Outcomes in Patients Initiated on Antiretroviral Therapy: A Prospective Cohort Study From Ethiopia. Drug safety. 2015;38(7):629–639. doi: 10.1007/s40264-015-0295-7. [DOI] [PubMed] [Google Scholar]
- 25.Lichtenstein KA, Armon C, Buchacz K, Chmiel JS, Moorman AC, Wood KC, et al. Initiation of antiretroviral therapy at CD4 cell counts >/=350 cells/mm3 does not increase incidence or risk of peripheral neuropathy, anemia, or renal insufficiency. Journal of acquired immune deficiency syndromes. 2008;47(1):27–35. doi: 10.1097/QAI.0b013e31815acacc. [DOI] [PubMed] [Google Scholar]
- 26.Prosperi MC, Fabbiani M, Fanti I, Zaccarelli M, Colafigli M, Mondi A, et al. Predictors of first-line antiretroviral therapy discontinuation due to drug-related adverse events in HIV-infected patients: a retrospective cohort study. BMC infectious diseases. 2012;12:296. doi: 10.1186/1471-2334-12-296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Hazell L, Shakir SA. Under-reporting of adverse drug reactions : a systematic review. Drug safety. 2006;29(5):385–396. doi: 10.2165/00002018-200629050-00003. [DOI] [PubMed] [Google Scholar]
- 28.Lopez-Gonzalez E, Herdeiro MT, Figueiras A. Determinants of under-reporting of adverse drug reactions: a systematic review. Drug safety. 2009;32(1):19–31. doi: 10.2165/00002018-200932010-00002. [DOI] [PubMed] [Google Scholar]
- 29.Montessori V, Press N, Harris M, Akagi L, Montaner JS. Adverse effects of antiretroviral therapy for HIV infection. CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne. 2004;170(2):229–238. [PMC free article] [PubMed] [Google Scholar]
- 30.Dean GL, Edwards SG, Ives NJ, Matthews G, Fox EF, Navaratne L, et al. Treatment of tuberculosis in HIV-infected persons in the era of highly active antiretroviral therapy. Aids. 2002;16(1):75–83. doi: 10.1097/00002030-200201040-00010. [DOI] [PubMed] [Google Scholar]