To assesses the safety and rationale of antibacterial fixed-dose combinations in the private sector in Latin America and determine the extent of their use.
To assesses the safety and rationale of antibacterial fixed-dose combinations in the private sector in Latin America and determine the extent of their use.
Analysis of FDCs was based on retail sales data for eight Latin American countries (Argentina, Brazil, Chile, Colombia, Mexico, Peru, Uruguay and Venezuela) between 1999 and 2009. FDCs were classified according to the pre-defined criteria. Use was expressed as daily defined doses per 1000 inhabitants per day (DDD/TID).
A total of 175 antibacterial FDCs contained a mean of 1.3 antibacterial substances and 3.2 other active substances. Thirty-seven (21%) FDCs were classified as unsafe, 124 (70%) as lacking sufficient evidence for efficacy and only 14 (9%) of all FDCs were considered rational, for example amoxicillin and clavulanic acid. Consumption of unsafe FDCs decreased by 0.011 DDD/TID (95% CI: −0.012 to −0.009) annually, from 0.173 DDD/TID in 1999 to 0.070 DDD/TID in 2009 (overall decrease, 59.7%). Consumption of FDCs lacking sufficient evidence decreased by 30.3% (−0.018 DDD/TID [95% CI: −0.028 to −0.008] annually), while use of rational FDCs increased by 17.1% (from 1.283 DDD/TID to 1.497 DDD/TID annually).
The majority of antibacterial FDCs in the private sector lack therapeutic benefit. Despite the decrease in the consumption of unsafe antibacterials and those lacking sufficient evidence, their use remains high and their marketing does not fit into strategies of prudent use of antibiotics to contain antibacterial resistance.
Evaluer la sécurité et la justification des combinaisons d'antibactériens à dose fixe (CADF) dans le secteur privé en Amérique latine et déterminer l’étendue de leur utilisation.
L'analyse des CADF a été basée sur les données des ventes en détail dans huit pays d'Amérique latine (Argentine, Brésil, Chili, Colombie, Mexique, Pérou, Uruguay et Venezuela) entre 1999 et 2009. Les CADF ont été classées selon des critères prédéfinis. L'utilisation a été exprimée en Doses Journalière Définies pour 1000 habitants par jour (DDD/TID).
Un total de 175 CADF contenaient en moyenne 1,3 substances antibactériennes et 3,2 autres substances actives. Trente-sept (21%) CADF ont été classées comme dangereuses, 124 (70%) comme manquant de preuves suffisantes d'efficacité et seulement 14 (9%) de toutes les CADF ont été considérées comme justifiées, ex: amoxicilline et acide clavulanique. La consommation de CADF dangereuses a diminué de 0.011 DDD/TID (IC95%: - 0.012 à – 0.009) chaque année, passant de 0.173 DDD/TID en 1999 à 0.070 DDD/TID en 2009 (diminution globale de 59.7%). La consommation de CADF non dotées de preuves suffisantes a diminué de 30.3% (−0.018 DDD/TID [IC95%: −0.028 à −0.008] par an), tandis que l'utilisation rationnelle des CADF a augmenté de 17.1% (passant de 1.283 DDD/TID à 1.497 DDD/TID par an).
La majorité des CADF dans le secteur privé manque de bénéfice thérapeutique. Malgré une baisse de la consommation d'antibiotiques à risque et de ceux manquant des preuves suffisantes, leur utilisation reste élevée et leur commercialisation ne rentre pas dans les stratégies d'utilisation prudente des antibiotiques pour endiguer la résistance aux antibactériens.
Evaluar la seguridad y el fundamento de combinaciones de dosis fijas (CDFs) de antibacterianos en el sector privado en América Latina y determinar la extensión de su uso.
El análisis de las CDFs se basó en datos de ventas al por menor de ocho países Latinoamericanos (Argentina, Brasil, Chile, Colombia, Méjico, Perú, Uruguay y Venezuela) entre 1999 y 2009. Las CDFs se clasificaron según criterios predefinidos. El uso se expresó como Dosis Diarias Definidas por cada 1000 habitantes por día (DDD/TID).
Un total de 175 CDFs de antibacterianos contenían una media de 1.3 sustancias antibacterianas y otras 3.2 sustancias activas. Treinta y siete (21%) CDFs fueron clasificadas como no seguras, 124 (70%) como faltándoles suficiente evidencia de eficacia, y solo un 14 (9%) de todas las CDFs fueron consideradas como fundamentadas, ej. amoxicilina y ácido clavulánico. El consumo de CDFs no seguras disminuyó en 0.011 DDD/TID (95% IC: − 0.012 a – 0.009) anualmente, de 0.173 DDD/TID en 1999 a 0.070 DDD/TID en 2009 (disminución total 59.7%). El consumo de CDFs con falta de evidencia disminuyó en un 30.3% (−0.018 DDD/TID [95% IC: −0.028 a −0.008] anualmente), mientras que el uso racional de CDFs aumentó en un 17.1% (de 1.283 DDD/TID a 1.497 DDD/TID anualmente).
A la mayoría de CDFs de antibacterianos en el sector privado les hace falta un beneficio terapéutico. A pesar de la caída en el consumo de antibacterianos no seguros así como de aquellos sin suficiente evidencia, su uso continúa siendo alto y su marketing no sigue estrategias de uso prudente de antibióticos que contengan el aumento de la resistencia a antibacterianos.
Antimicrobial drug resistance among common pathogens is rising rapidly with serious consequences for morbidity, mortality and healthcare costs. In Latin America, as in many other low- and middle-income countries that have a high burden of infectious diseases and limited healthcare budgets, this is of particular concern (Okeke et al. 2005). It is widely accepted that high usage of antibiotics is linked to the development of resistance (Goossens et al. 2005). International organisations have advocated the prudent use of antibacterials in the fight to curb the emergence of resistant bacteria. WHO launched a new package of interventions to improve use of antibiotics, stressing the urgency of action (Leung et al. 2011). The European Union funded a series of projects to monitor the use of antibiotics in Europe, for example, European Surveillance of Antimicrobial Consumption (ESAC) and Self-medication with Antibiotics and Resistance (SAR) (Goossens et al. 2005; Grigoryan et al. 2010). In the United States, an Interagency Task Force was created to develop and coordinate actions at federal level to combat resistance and promote prudent use of antibiotics (Center for Disease Control 2012). Only a few other countries have created national bodies to tackle the problem (Leung et al. 2011). Despite these efforts, improving antibiotic use is complex.
A fixed-dose combination (FDC) contains two or more active substances within a single pharmaceutical form of administration (European Medicines Agency 2009). The availability of such FDCs containing at least one antibacterial substance seems to be at odds with a strategy of prudent antibacterial use. In some cases, there may be improved adherence because of the convenience of taking fewer pills (European Medicines Agency 2009). In a few cases, there may be a pharmacological rationale of combinations of synergistically acting substances. The WHO List of Essential Medicines contains two pairs: amoxicillin and clavulanic acid (β-lactamase inhibitor) and trimethoprim and sulphamethoxazole (combination of two folic acid synthesis pathway inhibitors) (World Health Organization 2011). In addition, for the short-term therapy of eradication of Helicobacter pylori, a combination of two antibacterials with a proton pump inhibitor has been recommended (World Gastroenterology Organization 2010). But by and large, the rationality of many currently available FDCs containing antibacterial substances remains unclear (including combinations of an antibacterial substance with an analgesic or a mucolytic). Irrational use of antibacterial FDCs may also pose a risk to individual patients because of toxicity or lack of effects and thus cause unnecessary expense. Many irrational FDCs have been removed from the market in the United States and in Europe (Podolsky & Greene 2011), but to date, few such initiatives have been undertaken in low- and middle-income countries (Gautam & Aditya 2006). Studies in Latin America report that fixed-dose combinations (FDCs) of antibacterials are commonly used in the private sector (Gustafsson & Wide 1981; Wolff 1993; DURG-LA 1997) but the volume of their consumption over time and their rationality and safety have not been assessed in a systematic way. In the public sector in Latin America institutions, such as social security and Ministry of Health, only a very limited number of medicines are used within formularies. In contrast, the private sector market offers many pharmaceutical preparations than the public sector, among them antibacterial FDC. Recognising the extent of use of unsafe and irrational antibacterial FDCs in the private sector could be a first step in promoting appropriate use.
The aim of the present study was to assess the safety and rationale of antibacterial FDCs in the private sector in eight Latin American countries and determine the sales volume over time.
We retrieved private-sector sales data of all marketed antibacterial products for use in humans – either as single active substance or as fixed-dose combination (FDC) products – for systemic use (J01 ATC classification) from eight Latin American countries (Argentina, Brazil, Chile, Colombia, Mexico, Peru, Uruguay and Venezuela), representing 85% of the population in the region, between 1999 and 2009. Those countries were chosen because their pharmaceutical markets are the biggest in Latin America. These data were provided free of charge by IMS Health for this study (IMS-Health 2009). Data were based on information from manufacturers, retail wholesalers and, for some countries, pharmacies. The data represent the total national sales of all products, including antibacterials to retail sectors (sales in private pharmacies, private clinics and hospitals) and excluding those purchased by the public sector (government providers such as Ministry of Health or Social Security) as it relies in general on separate distribution channels in all eight countries studied.
Even though the health care systems of the eight countries under study are different, they have many things in common, in particular some fundamental organisational structures such as public and private pharmaceutical sector (PAHO, 2012). The public sector supply of medicines comprises the social security institutions and the Ministry of Health. Those institutions have their own pharmacies and dispensing outlets as well as their institutional formularies, which determine the medicines procured and prescribed. The institutional formularies are frequently reviewed by therapeutic committees and contain a very limited number of antibacterial FDCs, usually only those of the WHO Model List of Essential Medicines. The private sector includes all health institutions funded by non-governmental institutions and most are for profit. Within the private sector, the retail pharmacies or community pharmacies are providing access to medicines for the population. By and large for all eight countries, medicines consumption in the private sector represents out-of-pocket expenditure and only a small percentage private insurances (International Society for Pharmacoeconomics and Outcome Research, 2007).
We included all FDCs for systemic application, which contained at least one systemic antibacterial substance in the J01 group according to the WHO ATC classification (WHO Collaborating Centre for Drug Statistics Methodology 2011) as well as one or more other active substance(s) with or without antibacterial activity. Excluded are agents with antibacterial properties but not belonging to the J01 group, such as tuberculosis medication, which belong to the J04 group. Antibacterial FDCs were classified according to their main therapeutic indication based on manufacturer statements in printed and online physician desk references (DEF PLM for Mexico, Peru and Colombia) (Thomson PLM Mexico 2008a; Thomson PLM Peru 2008b; Thomson PLM Colombia 2010), textbooks (Parker et al. 2005; Sweetman 2009) and the scientific literature (accessed through Medline and EMBASE): Helicobacter pylori, gastrointestinal (except Helicobacter pylori), urogenital and respiratory infectious diseases. FDCs that could not be classified in any of the above categories were classified as ‘others’ including those that are used to treat sexually transmitted infections.
FDCs were then categorised as (A) ‘unsafe’, (B) ‘lacking evidence’ or (C) ‘rational’. We used a flow chart to guide classification based on international criteria taken from EMA (2009) and FDA (2011) guidelines on FDCs (Figure 1). We first determined whether active substances were considered sufficiently safe, alone or in combination, to be used in the proposed setting as part of an FDC. If not, we categorised the FDC as ‘unsafe’ (A). Second, we determined whether all active substances contributed to the claimed effect. We searched for published data (Medline, EMBASE) on a synergistic mechanism of action (pharmacodynamic or pharmacokinetic) that existed for the combined use of the active substances over and above the mere convenience of not having to take multiple tablets. If data were available, we classified the FDC as ‘rational’ (C); otherwise, it was classified as lacking evidence (B). All FDCs were classified in consensus by all co-authors.
Retail data for antibacterial substances (single active substances and FDCs) were converted into comparable units using the WHO concept of daily defined doses (DDD). A DDD is the assumed average maintenance dose per day for a drug used for its main indication in adults and is widely used in research on drug utilisation (WHO Collaborating Centre for Drug Statistics Methodology 2011). If the FDCs contained one systemic antibacterial substance (J01 group), we used the WHO-determined DDD of that particular substance. If the FDC contained more than one antibacterial substance, we used the WHO-determined DDD for the combination if available (e.g. the combination of trimethoprim and sulphamethoxazole [J01EE]) (WHO Collaborating Centre for Drug Statistics Methodology 2011). For all other FDCs containing more than one antibacterial substance, we used the sum of DDDs of the antibacterial substances with a J01 code. In one case, sulfathiazole, the ATC index did not provide a DDD for the J01 code and thus was excluded from the analysis of consumption. A request to assign a DDD was sent to the WHO Collaborating Centre. Eleven other combinations (with very low consumption according to the number of packages sold) were excluded in the analysis of sales volume because they lacked the information to calculate the DDDs per package (strength, pack size).
To allow cross-country comparison of antibacterials, we expressed the data as DDD per 1000 inhabitants per day (DDD/TID) where the total amount (in gram) consumed of each antibacterial substance per day was divided by the (calculated) DDD. This was then divided by the number of inhabitants of the respective countries multiplied by 1000. The number of inhabitants in the Latin American countries was based on the mid-year populations of the countries (Pan American Health Organization 2008). We calculated the DDD/TID for single antibacterials to compare that with the FDC data. We used linear regression to analyse whether consumption changed over time for all three FDC groups (‘unsafe’, ‘lacking evidence’ and ‘rational) and single antibacterials. We calculated the R square value for testing the fit; an R square value of 1 means a 100% fit between regression model and the data points. We also calculated a 95% confidence interval around the slope. If the 95% confidence interval for the slope excludes the value zero, it means that in 95% of the cases, the best fitted line would not be horizontal. All statistical analyses were conducted by using the statistical software package SPSS version 16 (SPSS Inc., Chicago, IL, USA).
Of 4487 different marketed antibacterial products for systemic use in the private sector in eight Latin American countries, 881 (19.6%) were identified as FDC products and the rest were products containing single antibacterials. Many of these products contained the same single or combination of active substances. Of the 881 FDC products, 175 were a unique composition of active substances. Of the 3606 single antibacterial products, 126 were unique antibacterial substances.
The 175 unique antibacterial FDCs contained a mean of 1.3 antibacterial substances (maximum four) and a mean of 3.2 active substances (maximum ten). Most commonly, combinations were classified as indicated for treating respiratory disease including cough and cold (n = 69), or gastrointestinal disorders such as diarrhoea (n = 36), followed by FDCs indicated for a variety of indications classified as ‘others’ (n = 41) and finally those for urogenital diseases (n = 25) (Table 1).
|Classification by indication||N||Mean no. antibacterials (JO1) (SD)||Mean no. active substances||Unsafe (A)||Lacking sufficient evidence (B)||Potentially rational (C)|
|Gastrointestinal diseases||36||1.2 (0.4)||3.9 (2.0)||8||Colistin + dicycloverine + diiodohydroxyquinoline + furazolidone||28||Pectin + phthalylsulfathiazole + streptomycin + sulfadiazine||0|
|Helicobacter Pylori eradication||4||1.8 (0∙5)||2.8 (0.5)||0||1||Amoxicillin + omeprazole||3||Amoxicillin + clarithromycin + omeprazole|
|Respiratory diseases||69||1.1 (0.4)||3.6 (1.6)||21||Methenamine + salicylic acid + theobromine||47||Amoxicillin + bromhexine||1||Amoxicillin + clavulanic acid|
|Urogenital diseases||25||1.5 (0∙.)||2.5 (0∙7)||4||Chloramphenicol + nitrofurantoin + phenazopyridine + sulfamethoxypyridazine;||19||Nalidixic acid + oxytetracycline + sulfamethoxypyridazine + phenazopyridine||2||Sulfadiazine + trimethoprim|
|Others||41||1.3 (0.6)||2.5 (1.2)||4||Oxyphenbutazone + papain + tetracycline||29||Metronidazole + pyridoxine + riboflavin + thiamine||8||Sulphamethoxazole + trimethoprim|
|Total||175||1.3 (0.5)||3.2 (1.6)||37||124||14|
Thirty-seven (21%) of all 175 unique FDCs contained unsafe active substances (= category A) (Table 2). Unsafe substances were antibacterials such as diiodohydroxyquinoline (neurotoxicity including permanent loss of vision) and thiamphenicol/chloramphenicol that were considered suitable only after careful weighing of its antibacterial properties vs. its serious haematological adverse effects rendering these drugs unsuitable for use in an FDC (23). A few antibacterial FDCs were considered unsafe because of the non-antibacterial active substances they contained, for instance theophylline, which should not be used in combination due to its narrow therapeutic index.
|Unsafe substance||No. FDCs||Example of FDC||Reason|
|Chloramphenicol+ as fixed-dose combination||8||Chloramphenicol + chlorphenamine + guaifenesin + iodine + oxeladin + phenylephrine + prednisolone||Chloramphenicol should not be used in fixed-dose combinations. Chloramphenicol has a place in therapy for very defined indications and should only be used under medical supervision and after careful weighing of its antibacterial benefits against its potentially severe haematological adverse effects, which include reversible bone marrow suppression and irreversible aplastic anaemia (Yunis 1988)|
|Thiamphenicol as fixed-dose combination||2||Acetylcysteine + edetic acid + thiamphenicol||Thiamphenicol should not be used in fixed-dose combinations. It has a similar side-effect profile as chloramphenicol (Yunis 1988)|
|Diiodohydroxyquinoline||2||Diiodohydroxyquinoline + homatropine + kaolin + pectin + sulfathiazole||Severe neurotoxicity including peripheral neuropathy, myelopathy and optic atrophy leading to permanent loss of vision (Oakley 1973)|
|Metamizole sodium+||16||Ampicillin + bromhexine + guaifenesin + lidocaine + metamizole sodium||Metamizole should not be used as fixed-dose combination with antibiotics. Metamizole might have a place in therapy as a single substance but its benefits need to be carefully weighed against its severe haematological adverse effects: agranulocytosis (Ibanez et al. 2005)|
|Oxyphenbutazone||3||Tetracycline + papain + oxyphenbutazone||Severe haematological adverse effects: aplastic anaemia (Drug and Therapeutic Bulletin 1984)|
|Atropine/atropa belladonna||3||Atropine + bismuth salt+ charcoal + menadione + neomycin + papaverine + phthalylsulfathiazole + riboflavin + thiamine||Anticholinergic adverse effects (Parker et al. 2005; Sweetman 2009)|
|Ephedrine||1||Cephaelis ipecacuanha + ephedrine + oxytetracycline||Cardiac and central nervous toxicity including hypertension, arrhythmia, stroke (Centers for Disease Control and Prevention 1996; Haller & Benowitz 2000)|
|Theobromine/theophylline||3||Methenamine + salicylic acid + theobromine/theophylline||Cardiac and central nervous toxicity (Parker et al. 2005; Sweetman 2009)|
One hundred and twenty-four (70%) FDCs lacked evidence (= category B) for the combined use of active substances (Table 1). These combinations were mainly found in FDCs intended for treatment of respiratory diseases or gastrointestinal diseases. Examples include the combination of bromhexine or ambroxol and amoxicillin, and the combination of loperamide and neomycin. Many combinations contained non-steroidal anti-inflammatory drugs (NSAIDs) for the treatment of fever or pain associated with the infection. Phenazopyridine was the most frequent analgesic substance in FDCs that included antibacterials such as nalidixic acid or sulphonamides. Other examples were combinations of antimicrobial agents for which there was no evidence for superior efficacy compared to the – first mentioned – single antibacterials, for example, metronidazole and spiramycin, ampicillin and dicloxacillin or benzoic acid and methenamine. Furthermore, five FDCs included substances that enhanced the pharmacokinetic or pharmacodynamic properties of the antibacterial substance. Examples included the combination of amoxicillin and probenecid or doxycycline and lysozyme, however, with no clear evidence for superior clinical efficacy compared to the single antibacterials.
Finally, 14 (8%) of all FDCs were classified as rational (= category C). These were combinations of two antibacterial substances and a proton pump inhibitor to treat gastrointestinal infections with Helicobacter pylori, combinations of amoxicillin and clavulanic acid/pivsulbactam, or combinations of trimethoprim and a sulphonamide.
During the study period, antibiotic consumption in the retail sector decreased between 1999 and 2002 by −0.254 DDD/TID (95% CI: −0.412 to −0.078; P = 0.024; R2 = 0.928) per year and increased by +0.266 DDD/TID (95% CI: 0.243–0.288; P = 0.000; R2 = 0.994) between 2002 and 2009. This resulted in an overall increase of 12.0% within the study period. This overall growth was explained by increasing consumption of single antimicrobials + 0.155 DDD/TID (95% CI: 0.095–0.215; P = 0.000; R2 = 0.769) annually, from 6.76 DDD/TID in 1999 to 7.95 DDD/TID in 2009.
In 1999, 7.8% of DDD/TID were unsafe FDCs, 35.2% lacked evidence and 57.0% were rational FDCs. In 2009, 3.3% were unsafe, 25.9% lacked evidence and 70.8% were rational FDCs. While the consumption of antibiotic FDCs remained largely constant over the study period −0.007 DDD/TID (CI 95%: −0.028 to +0.013; P = 0.421; R2 = −0.030) (see also Figure 2), its percentage out of total antibiotics decreased slightly [−0.4% annually; 95% CI: −0.4 to −0.5%], from 25% (2.22 DDD/TID) in 1999 to 21% (2.10 DDD/TID) of total antibacterial consumption in 2009 (Figure 2). As seen in Figure 2, FDCs lacking evidence (B) decreased by −0.018 DDD/TID (95% CI: −0.028 to −0.008; P = 0.002; R2 = 0.620) annually, from 0.766 DDD/TID in 1999 to 0.534 DDD/TID in 2009 (overall decrease 30.3%). Unsafe FDCs (A) decreased significantly by 0.011 DDD/TID (95% CI: −0.012 to −0.009; P = 0.000; R2 = 0.973) annually, from 0.173 DDD/TID in 1999 to 0.070 DDD/TID in 2009 (overall decrease, 59.7%) (Figure 3), which is largely explained by reduction in FDCs containing metronidazole and diiodohydroxyquinoline (Figure 3). In contrast, rational FDCs remained largely constant between 1999 and 2004 (−0.007 DDD/TID; 95% CI: −0.019 to 0.005; P = 0.185; R2 = 0.238) but afterwards increased by 0.047 DDD/TID (95% CI: 0.027–0.0681; P = 0.005; R2 = 0.929) annually, from 1.283 DDD/TID in 2005 to 1.497 DDD/TID in 2009 (overall increase, 17.1%) (Figure 4). Amoxicillin/clavulanic acid and trimethoprim/sulphamethoxazole (co-trimoxazole) FDCs comprised 93.1% of the rational FDCs, with the share of amoxicillin/clavulanic acid growing from 26.7 to 64.3% in 2009 at the cost of co-trimoxazole sales (Figure 4).
When analysing consumption by country, FDCs ‘lacking evidence’ were sold in all countries. In 2009, this was out of total FDC consumption, highest for Mexico and Argentina (33.4% and 27.3%, respectively), medium for Peru (20.2%), Uruguay (16.3%), Colombia (16.3%) and Brazil (11.5%), and lowest for Chile (2%) and Venezuela (2%). Consumption for unsafe FDC was highest in Uruguay (5%) and Brazil (2%) followed by Argentina (0.7%), Peru (0.7%) and Mexico (0.2%); no sales were recorded for Chile, Colombia and Venezuela.
To our knowledge, this is the first longitudinal analysis of rationale and sales of antibacterial FDCs in the private health sector of Latin America, where in 2009, 20% of total antibiotic use was in the form of FDCs. And although there was a 60% decrease in unsafe antibacterial FDCs in the study period, sales of unsafe FDCs or those lacking evidence remains high, with more than a quarter (29%) of the sold antibacterial FDCs lacking evidence of therapeutic rationale.
The total average consumption of antibiotics for the eight Latin American countries in 2009 is comparable to the total outpatient consumption that was reported in 2002 for some of the lowest-use European countries (Goossens et al. 2005). However, this comparison must be made with some caution as our data underestimate total antibiotic use as we do not have data for the public sector (Wirtz et al. 2010). FDCs that lack evidence were sold in all countries studied, while in three of eight, no sales of unsafe FDCs were registered in 2009. This study focuses on the development in the whole region and further work should study the situation per country in more detail.
The decrease in consumption of FDCs that lack evidence is a positive development, and the trend of consumption over time suggests that it may continue to decrease until reaching a steady state. Substances such as diiodohydroxyquinoline (New Zealand Medicines & Medical Devices Safety Authority 1999) have become obsolete in many countries even when used as single antimicrobials because of their negative benefit to risk ratio resulting at best in a very narrow therapeutic margin. Chloramphenicol should not be used as a combination product; if used at all, it should be initiated as single product allowing flexible dosing schemes after careful consideration of an individual's benefit to risk ratio and with proper monitoring (Yunis 1988). An FDC containing various other substances in addition to chloramphenicol does not fit in such a therapeutic approach. We also identified seven other active substances that we considered to be unsafe for use in an FDC. For example, theophylline has a place in therapy, but due to serious adverse effects and a narrow therapeutic window, it is not a suitable candidate to be included in combination with antibiotics to treat infections (Food & Drug Administration 2011). All those combinations are not in line with the WHO key recommendation to contain antimicrobial resistance by ensuring marketing of safe and effective products (WHO 2001).
We found a large number of combinations lacking evidence for which the sales also decreased. We decided to use a definition that would potentially result in under- rather than overestimation. For example, the majority of FDC has an approved indication for cough- and cold-related symptoms or gastrointestinal diseases such as diarrhoea, conditions for which antibiotics generally are not indicated. The risk of adverse events and the potential to create resistance likely outweighs the benefit of those medicines and are therefore unsafe. Some of the FDCs were difficult to classify as there was a paucity of information on their effects and toxicity. Examples include nifuroxazide and diiodohydroxyquinoline (Oakley 1973). Other FDCs contained substances that are considered obsolete because there are better alternatives. Examples include the first quinolone antibiotics such as nalidixic acid or pipemidic acid (Ball 2003). Information by the manufacturers varied in style, length and reliability. Many of the FDC identified in our study do not meet the criteria of the European Medicines Agency (2009), which guided our assessment in terms of indications of well-recognised disease states or two closely related diseases. Furthermore, there is little information as to whether an FDC should be used as first-line, second-line or substitution therapy, let alone whether it may (not) be used in children. The argument of convenience and enhancing compliance –sometimes used by the pharmaceutical industry – is hardly defensible because of the lack of evidence for efficacy of many of these FDCs. At first glance antibacterial FDCs containing analgesics or non-steroidal anti-inflammatory drugs may have some rationale due to infection-associated fever, inflammation and pain. However, fever and pain usually subside quickly after initiation of antibiotic therapy, after which the analgesic component of the FDC becomes superfluous.
In Latin America, as is likely in many other low- and middle-income countries, the availability of unsafe and irrational FDCs is of particular concern: products can be accessed and are regularly used without medical supervision because unauthorised OTC sales of prescription-only-medicines is common (Organización Panamericana de la Salud 2004). The concern is shared by the United States, where a number of citizens cross the border to Mexico to obtain medicines, including antibiotics (US Department of State 2012). However, rational, FDCs or single antibiotic products when used inappropriately may also be harmful for an individual patient or promote development of resistance to antibiotics.
We must consider the possible limitations of our data. The data represent the industry standard for estimating national retail sales volume of medications. Similar data have been widely used in studies of antibiotic utilisation patterns (New Zealand Medicines & Medical Devices Safety Authority 1999), among them the ESAC project (Goossens et al. 2005). Changes in the way the data are collected over time could have affected our results. However, the data provider (IMS-Health 2009) is carrying out frequent validation of wholesalers' and retail sales data with production reports from manufactures. For all eight countries, the percentage of retail sales data captured is between 90 and 98%. The ratio between public and private sector varies in each of the eight countries studied. Our objective was to study retail sales volume, which represents for five of eight countries (Argentina, Brazil, Chile, Uruguay and Venezuela) more than 80% of the total pharmaceutical volume consumed in the country. However, in the Latin American setting, as in many other low- or middle-income countries, independent and/or representative national drug databases are virtually non-existent so that external validation of the data is not possible at present. Using linear regression for consumption trend analysis shows a good fit of the data. There were insufficient data points for nonlinear analysis.
Availability and consumption of unsafe antibacterial FDC or those with insufficient evidence in Latin America remains worryingly high. Given the paucity of basic information on safety and effectiveness of many substances contained in the FDC, they do not fit into strategies of prudent use of antibiotics to contain antibacterial resistance. In our study, we used an easy-to-use tool to assess the safety and rationale of antibacterial FDCs that could assist policy makers and regulators – not just in Latin America – to reassess the locally available therapeutic arsenal and withdraw those FDC considered unsafe or with insufficient evidence.
We would like to thank Yared Santa Ana Tellez for her support in the data analysis.