Antibiotic misuse and the emergence of resistance

Despite the link between use and resistance, the appropriate use of antibiotics is one of humankind's most essential weapons against disease. So, interventions need to target inappropriate patterns of use, specifically those that have contributed most significantly to the development of resistance. In this section, we examine antibiotic misuse and explore its root causes.

Common colds - acute inflammatory changes, mostly mediated by viruses, anywhere along the continuum of the upper respiratory tract - lead to 110 million outpatient visits and cost an estimated $40 billion annually in the US. On average, common colds afflict each US adult 2.2 times per year, and child 3 times per year (Fendrick et al., 2003). Despite the fact that viruses cause >90 percent of these maladies, physicians regularly prescribe antibacterial agents to treat acute upper respiratory tract infections (URIs).

Researchers used the National Ambulatory Medical Care Survey (NAMCS), an annual sampling of reasons people seek outpatient medical care, to approximate the quantity and cost of antibiotic use for URIs. NAMCS data revealed that in 1998, 84 million office visits for URIs led to 45 million antibiotic prescriptions. Using historical data to extrapolate bacterial infection rates, they concluded that 55 percent of the antibiotics were prescribed inappropriately for

ZD DDD/1000 inhabitants/day % amoxicillin resistant E. coli

Newfoundland Netherlands Greece

ZD DDD/Km2/day

% amoxicillin resistant E. coli isolates

Newfoundland Netherlands Greece

Figure 9.8 Prevalence of amoxicillin-resistant E. coli from three cities, plotted against aminopenicillin consumption (DDD/1000 per day) and consumption as a function of population density (DDD/km2 per day). Adapted from Bruinsma et al. (2003) with permission.

Newfoundland Netherlands Greece

Figure 9.8 Prevalence of amoxicillin-resistant E. coli from three cities, plotted against aminopenicillin consumption (DDD/1000 per day) and consumption as a function of population density (DDD/km2 per day). Adapted from Bruinsma et al. (2003) with permission.

illnesses of viral origin. Antibiotics prescribed for URIs led to costs of $1.32 billion, of which an estimated $726 million was for unneeded antibiotics (Gonzales etal, 2001; Figure 9.9). More recently, researchers estimated that, annually, $1.1 billion was being spent on 41 million unnecessary antibiotic prescriptions for viral respiratory tract infections (Fendrick et al., 2003).

Similar misuse of antibacterial agents for URIs has been documented in Europe. A study of 2899 acute respiratory infections from 10 Spanish hospitals between 1994 and1995 showed that antibiotics had been prescribed for 83 percent; 41 percent of the prescriptions were considered inappropriate based on use for diseases of likely viral origin (Ochoa et al, 2000). National health survey data demonstrated similar trends in France, where the proportion of acute respiratory tract infections of presumed viral etiology treated with antibiotics increased by 86 percent for children and 115 percent for adults between 1981 through 1992 (Guillemot et al, 1997).

One may postulate that treating URIs of likely viral etiology with antibiotics may reduce rates of secondary bacterial infections; however, this view is unproven. Studies looking at the efficacy of antibiotic therapy for specific manifestations of URIs, such as bronchitis, sinusitis, and pharyngitis, consistently have failed to demonstrate a beneficial treatment effect. Fahey and colleagues, for instance, analyzed eight randomized, placebo-controlled trials, with greater than 300 patients in each arm, looking at the role of antibiotics versus placebo for acute cough-related illnesses in adults. They found no significant differences in proportion of patients with resolution of cough or improvement in symptoms by 7-11 days (Fahey et al, 1998).

25,000

□ Office visits

□ Antibiotic prescription ■ Bacterial prevalence o 15,000 — 0 0

5000

□ Office visits

□ Antibiotic prescription ■ Bacterial prevalence

URI Otiis Media Sinusitis Pharyngitis Bronchitis

Figure 9.9 Primary care office visits and antibiotic prescriptions for acute respiratory illnesses in the United States. Reproduced from Gonzalez et al. (2001), with permission.

URI Otiis Media Sinusitis Pharyngitis Bronchitis

Figure 9.9 Primary care office visits and antibiotic prescriptions for acute respiratory illnesses in the United States. Reproduced from Gonzalez et al. (2001), with permission.

The ecologic impact of inappropriate antibiotic prescribing for URIs may be illustrated most readily by examining the association between antibiotic use and resistance for a common, community-acquired bacterial isolate such as S. pneumoniae. This link has been demonstrated for both asymptomatic carriage and invasive pneumococcal disease, with patient- and population-level analyses. To evaluate this issue, general practitioners in Canberra, Australia, took nasal swab cultures and detailed antibiotic-use histories from the first 15 children, aged six years and under, who entered their practice on four separate occasions between 1997 and 1999. During the four sampling periods they collected a total of 1502 swabs, 631 (42 percent) of which grew S. pneumoniae; 14 percent were penicillin-resistant S. pneumoniae (PRSP). Utilizing a multivariate regression model, the authors identified beta-lactam use in the two months before swab collection as a key risk for carriage of PRSP. Children who had received both penicillin and a cephalosporin in the preceding two months were at even greater risk. Notably, any antibiotic use upto six months before specimen collection conferred an increased risk for PRSP carriage (Nasrin etal, 2002; Figure 9.10). A similar study correlated nasal carriage rates of PRSP for children under seven years of age in Iceland with individual, as well as regional, antibiotic use. Children from the region with the highest overall antibiotic consumption (23 vs 9.6 DDD per 1000 children per day in the lowest use region) had an odds ratio of 20.3 for PRSP carriage. Additionally, children who had received a beta-lactam antibiotic in the previous year had a 6.75 odds ratio for PRSP carriage (Vilhjalmur et al, 1996).

In addition to promoting PRSP carriage, outpatient antibiotic use has also been linked to invasive PRSP disease. In a retrospective cohort study, researchers reviewed records from 374 patients with invasive pneumococcal disease from

(257 isolates) (64 isolates) (74 isolates) (61 isolates)

Days of g lactam use in the six months before the swab

Figure 9.10 Penicillin resistance in pneumococcus isolates from children, and beta-lactam use in the six months before swab collection. Taken from Nasrin et al. (2002).

DDD beta-lactam antibiotics/1000

Figure 9.11 The log odds of resistance to penicillin among invasive isolates of Sreptococcus neumoniae (ln (R/{1/R})) is regressed against outpatient sales of beta-lactam antibiotics in 11 European countries; resistance data are from 1998-99 and antibiotic sales data are from 1997. DDD, defined daily dose; BE, Belgium; DE, Germany; FL, Finland; EI, Ireland; IT, Italy; LU, Luxembourg; NL, Netherlands; PT, Portugal; ES, Spain; SE, Sweden; UK, United Kingdom. From Bronzwaer et al. (2002).

DDD beta-lactam antibiotics/1000

Figure 9.11 The log odds of resistance to penicillin among invasive isolates of Sreptococcus neumoniae (ln (R/{1/R})) is regressed against outpatient sales of beta-lactam antibiotics in 11 European countries; resistance data are from 1998-99 and antibiotic sales data are from 1997. DDD, defined daily dose; BE, Belgium; DE, Germany; FL, Finland; EI, Ireland; IT, Italy; LU, Luxembourg; NL, Netherlands; PT, Portugal; ES, Spain; SE, Sweden; UK, United Kingdom. From Bronzwaer et al. (2002).

five Spanish hospitals over a five-year period; 24 percent of the episodes were due to PRSP. Beta-lactam exposure in the prior three months was the only risk factor for penicillin resistance identified by multivariate analysis (Nava et al, 1994). These data have been corroborated on a national level. EARSS collected national rates of penicillin non-susceptible S. pneumoniae (PNSP) from patients with invasive disease and outpatient antibiotic consumption data from 11 countries. They found a convincing correlation between beta-lactam consumption and national rates of PNSP infection (Bronzwaer et al., 2002; Figure 9.11).

The recognition that inappropriate antibiotic prescribing for URIs has contributed to the global emergence of antibiotic resistance begs the simple question: what drives physicians to prescribe antibacterial agents for diseases of likely viral etiology? Although 97 percent of 350-plus Georgia (USA) practitioners surveyed about their antibiotic-prescribing practices agreed that antibiotic overuse represents a major factor contributing to resistance, 42 percent admitted to prescribing antibiotics for the common cold (Watson et al, 1999). In a survey designed to elicit information about prescribing practices of nearly 500 Massachusetts primary care providers, 62 percent identified diagnostic uncertainty as a major reason for prescribing antibiotics for URIs (APUA, 1999). In the face of diagnostic uncertainty, physicians may feel obligated to err on the side of benefiting the patient by treating a possible bacterial infection, at the expense of the more distant and community-level problem of antibiotic resistance.

Patient or parent expectations also heavily influence physicians' antibiotic prescribing practices; 59 percent of 500 Massachusetts physicians surveyed indicated that patient requests increased their antibiotic prescribing (APUA, 1999). In a questionnaire-based study of over 600 pediatricians, 40 percent of respondents indicated that, on 10 or more occasions in the month before the survey, parents had requested an antibiotic when the physician did not feel it was indicated; 48 percent of respondents reported that parents always or most of the time pressured them to prescribe antibiotics; and one-third reported that they frequently complied with these parental requests. In this study, parental pressure, more than concerns about legal liability or practice efficiency, contributed most strongly to inappropriate antibiotic use (Bauchner et al., 1999).

In a review of prescribing patterns in 15 practices, researchers found that 60 percent of the adult patients expected an antibiotic for a URI. Patients were more likely to expect an antibiotic if they felt that they had benefited previously from antibiotics for similar symptoms. In the same study, physicians believed that 62 percent of their patients expected antibiotics, and prescribed antibiotics to these patients more frequently (Dosh et al., 2000). In a phone survey of over 5000 patients from nine countries, 11 percent admitted to exaggerating URI symptoms to get antibiotics from their physicians (Perchere, 2001). In a US survey, 32 percent of 12,000-plus people believed that taking antibiotics for colds prevented more serious illness, and 48 percent expected a prescription for antibiotics if they felt ill enough to seek medical attention (Eng et al., 2003).

In addition to patient expectations, economic factors may also figure largely into a physician's decision to prescribe antibiotics for URIs. Pediatricians may prescribe antibiotics in an attempt to limit working parents' need for return visits (Pichichero, 1999) - but in fact an antibiotic prescription may "legitimize" the office visit for URI symptoms, thereby increasing the likelihood of an office visit and antibiotic expectations for the child's next cold (Watson et al., 1999).

The details surrounding a physician's remuneration also may affect prescribing practices. Though all physicians in Canada receive payment from the single-payer government health system, some are salaried and some are paid on a fee-forservice basis. A study of antibiotic prescribing found that both fee-for-service payment and greater volume of patients strongly correlated with higher antibiotic prescription rates (Hutchinson and Foley, 1999). Additionally, in a managed-care setting that awards physicians based on patient satisfaction surveys, physicians may prescribe out of fear of dissatisfying their patients (Schwartz et al., 1998).

Some of the factors discussed above may be operating on a national level, leading to dramatic disparities in prescribing practices in the neighboring countries of France and Germany. In 1998, 7 percent and 53 percent of the S. pneumoniae strains from Germany and France, respectively, were PNSP. This difference parallels the 8 percent versus 48 percent outpatient visits for colds that culminated in antibiotic prescriptions in Germany compared with France. Factors that may drive these differences include increased pressure to prescribe among French patients and increased utilization of day care by the French, whose households are more likely to have two working parents. Macro-economic forces also may drive the disparity in prescribing practices. Retail drug prices in France are among the lowest in all of Europe - 32 percent less than US prices -whereas German prices are 24 percent higher than US prices (56 percent higher than French prices). Low drug prices likely have contributed to France's ranking highest in per capita outpatient antibiotic consumption in Europe. Evidence also suggests that low drug prices in France encourage use of newer, broad-spectrum agents, whereas Germans more often rely on less expensive, older, generic antibiotics (Harbarth et al, 2002b).

Antibiotics sold over the counter (OTC), without a physician's prescription, also account for a significant portion of antibiotic misuse, especially in developing countries, whose inhabitants may have limited access to physician-guided medical care. In addition to limited access to physicians, self-medication with antibiotics may be motivated by the desire to save money, the perceived need for urgent treatment of a suspected bacterial illness, or the desire to maintain privacy regarding potentially embarrassing symptoms. Varying legal and cultural forces shape pharmacy prescribing practices in different countries, including willingness to give advice on antibiotic use or to refill old prescriptions without a physician's approval (Radyowijati and Haak, 2002).

In many places, antibiotics may be sold without any physician advice. Of 49 Greek pharmacists presented with a fictionalized acute rhinosinusitis patient, 86 percent offered antibiotics, most often broad spectrum (Conotpoulos-Ioannidis et al, 2001). A similar study that presented fictional scenarios to 100 pharmacies in Katmandu, Nepal, found that all retail pharmacists engaged in diagnostic and therapeutic behavior beyond the scope of their training; 97 percent suggested inappropriate antibiotics for diarrhea, while many (56 percent) failed to suggest oral rehydration therapy and only 3 percent recommended seeking a physician's care (Wachter et al, 1999).

0 0

Post a comment