level and the use of antibiotics

PATIENTS, MATERIALS AND METHODS

The study was conducted from 1 March 2008-28 February 2009 at a primary care facility in

São Paulo, Southeast Region of Brazil. Children of less than 12 years of age were

attended by paediatricians and the clinical diagnosis was coded using the International

Classification of Disease, 10th Revision. All children seen on Tuesday's and Friday's

mornings with the diagnosis of common cold (J00) or acute upper respiratory infections

of multiple and unspecified sites (J06) with symptoms starting in the last five days

were included in the study sample. Patients with any associated codes for sinusitis,

otitis media, bacterial pharyngitis, pneumonia or who had underlying chronic heart or

lung disease or any other chronic health problem or immune disorder were excluded.

Patients who had used antibiotics in the last five days were also excluded. The study

was approved by the Ethical Committee of São Paulo Federal University (CEP 0670/08) and

written informed consent was obtained from parents or guardians before enrollment of

each patient.

An independent paediatrician who was not responsible for any clinical intervention

filled out a standardised case report form that included information on demographic

characteristics and clinical syndrome for each enrolled patient. Data of follow-up were

obtained in medical records of subsequent office visits and also by regular telephone

calls to the parents every two or three days. All data were entered on the form until

the resolution of the respiratory illness. Data about medical prescription of

antibiotics, symptoms on the day of antibiotic prescription, laboratory tests and X-ray

exams reports were recorded on the form. The use of antibiotics was considered

appropriate for all patients with signs and symptoms of secondary bacterial infection

diagnosed as AOM, bacterial sinusitis or pneumonia with evidence of parenchymal

infiltrates in the chest X-ray exam. For patients with the diagnosis of sinusitis, the

use of antibiotics was considered appropriate in cases of symptoms of rhinosinusitis and

cough without improvement for more than 10-14 days or more severe upper respiratory

tract signs and symptoms, i.e., fever ≥ 39ºC, facial swelling, facial pain (Dowell et

al. 1998a).

Laboratory tests - A sample of nasopharyngeal wash was obtained from

each patient at the enrollment visit. The median time from the onset of symptoms to the

collection of nasal washings was three days (1-5 days). Samples were collected and

processed using previously published procedures (Carraro

et al. 2007) and immediately transported to the laboratory for respiratory

virus testing.

Respiratory samples were tested first with the commercial panel SimulFluor Respiratory

Screen (Chemicon International, USA) for detection of seven respiratory viruses:

respiratory syncytial virus, influenza A and B viruses, parainfluenza virus types 1-3

and adenovirus, as described previously (Bellei et al.

2007, Carraro et al. 2007). Original

specimens were then stored at -80°C for further analysis and as polymerase chain

reaction (PCR) tests became available in our laboratory, the viral RNA and DNA were

extracted from each sample using QIAamp Viral RNA and DNA extraction Kit (Qiagen, USA)

according to the manufacturer's instructions. Nucleic acid extracts were tested by PCR

for the detection of respiratory syncytial virus, influenza A and B viruses, adenovirus,

rhinovirus, metapneumovirus, bocavirus, coronavirus and enterovirus, according to

published methods (Allard et al. 2001, Savolainen et al. 2002, Erdman et al. 2003,Falsey et al.

2003, Allander et al. 2005, Brittain-Long et al. 2008, Vijgen et al. 2008, CDC

2009).

Statistical analysis - Statistical analysis was performed using SPSS

20.0 and Stata 12 statistical softwares and the statistical significance level for all

tests was 5%. The existence of associations between two categorical variables was

verified using the chi-square test or the Fisher's exact test depending on the sample

size. The Student's t test was used to perform comparisons of averages

between two groups of data in cases of independent samples. To compare more than two

averages, the analysis of variances (ANOVA) was used. Normal distribution of data was

previously verified using Kolmogorov-Smirnov test. Nonparametric Mann-Whitney

U and Kruskal-Wallis tests were used to compare the averages of

samples that did not meet the normality assumption. If differences were found between

averages, multiple comparisons were performed to localise such differences. For the

pairwise comparisons of groups, adjustments in the descriptive levels were made so that

the overall significance was 5%.

RESULTS

During the whole study, 3,282 medical consultations were carried out, 29% (955/3282) of

which were coded as common cold or acute upper respiratory infections of multiple and

unspecified sites. Of these 955 common cold cases, a sample of 134 patients who met the

inclusion criteria was obtained, median age 2.9 years (0.1-11.2 y), 49% male. The most

frequent symptoms were coryza (91.8%, 123/134), cough (90.3%, 121/134), fever (56%,

75/134) and wheezing (46.3%, 62/134). Respiratory viruses were detected in 73.9%

(99/134) of nasopharyngeal wash samples (1 sample per patient) with a coinfection rate

of 30.3% (30/99). The laboratory tests findings are described in Table I.

Overall, the antibiotic prescription rate was 39.6% (53/134), among which 60.4% (32/53)

was amoxicillin, 22.6% (12/53) macrolides, 9.4% (5/53) cephalosporins and 7.6% (4/53)

amoxicillin plus sulbactam (Table II). Of 53

patients who received antibiotics during the follow-up, only 30.2% (16/53) received them

judiciously and the other 69.8% (37/53) received them inappropriately. Among these 37

cases with inappropriate use, the clinical justifications for prescription of

antibiotics were: in 37.7% (20/53) to treat nasal or postnasal discharge during the

first week of common cold symptoms in patients without fever, in 18.9% (10/53) to treat

persistence of cough during the first week of symptoms, in 11.3% (6/53) to treat common

cold and in 1.9% (1/53) to treat wheezing symptom.

Among 75 children who had fever at the onset of symptoms, 45.3% (34/75) were prescribed

antibiotics whereas 32.2% (19/59) of those who did not have fever at the onset of

symptoms received antibiotics. Thus, there was no difference in the proportion of

antibiotic prescriptions between children who had fever at the onset of symptoms and

those who did not (p = 0.123).

Of a total of 53 children who received antibiotics, 34 presented fever at onset of

symptoms and of these, 29.4% (10/34) received judicious prescription of antibiotics. Of

the remaining 19 children who did not have fever at the onset of common cold symptoms

and received antibiotics, 31.6% (6/19) were prescribed antibiotics judiciously. Thus,

there was also no difference for judicious prescription of antibiotics between children

with fever and those without fever at the onset of common cold symptoms (p = 0.869).

The average time to the resolution of symptoms of children with signs of secondary

bacterial infection was of 16.7 days and, within this group, all children received

antibiotics. Among children with no signs of bacterial infection, the average time to

the resolution of symptoms was 8.9 days for the group that received antibiotics and 7.0

days for the group that did not.

Among patients with respiratory virus monoinfection, all patients with influenza

received antibiotics inappropriately (10/10), whereas those with respiratory syncytial

virus were prescribed antibiotics inappropriately in 60% (3/5) and those with rhinovirus

were prescribed antibiotics inappropriately in 44.4% of cases (5/9) (p = 0.016). Also,

of seven patients coinfected with influenza, 71.4% (5/7) received antibiotics

inappropriately, as showed in Table II. None of

the patients were vaccinated against influenza by the time of the study.

DISCUSSION

Inappropriate use of antibiotics for viral common colds is an important problem

worldwide but, until now, we did not have clinical studies addressing this problem in

Brazil. During the clinical course of a common cold, secondary bacterial infections may

occur, however, antimicrobials are frequently used inappropriately for events that are

normal during the clinical course of a viral infection (Dowell et al. 1998b). In our

study, this practice was particularly evident in cases of viral rhinosinusitis, since

37.7% (20/53) of patients received antibiotics for the presence of nasal or postnasal

discharge during the first week of common cold symptoms without fever. During childhood,

20-40% of common colds can be complicated by AOM (Wald

et al. 1991, Chonmaitree et al. 2008)

and although current recommendations on management of AOM encourage the initial

observation of nonsevere AOM cases in selected children (i.e., “watchful waiting”)

(AAP 2004), there are still controversies

around the diagnosis and management of AOM (Hoberman et

al. 2011,Shaikh et al. 2011). In

routine practice, the over diagnosis of AOM is frequent but, considering that there are

difficulties in confirming the diagnosis, in our study we considered all indications of

antibiotic for AOM as adequate regardless of the patients' age, severity of symptoms and

certainty of diagnosis. We believe that rates of inappropriate use of antibiotics could

be even greater if stringent diagnostic criteria were applied for AOM diagnosis.

Amoxicillin was the first choice for antibiotic therapy accounting for 60.4% (32/53) of

all prescriptions but, on the other hand, the second most prescribed antibiotics were

macrolides (22.6%, 12/53) which is not in accordance with current guidelines (Bradley et al. 2011). Predominantly in the paediatric

outpatient population, the growing emergence of bacterial resistance to macrolides is

related to delayed cure and is responsible for a great proportion of costs currently

associated with antibiotic resistant pneumococcal infections (Kawai et al. 2012, Reynolds et al.

2014). Continuing medical education programs aimed not only to reduce the

inappropriate use of antibiotics, but also to teach the best practices of antibiotic

therapy could be of great help to improve this scenario.

Diagnostic uncertainty, which is defined as the difficulty in distinguishing a

self-limited viral infection from a bacterial infection requiring antibiotic therapy is

identified by researchers as a factor that contributes for antibiotic overuse (Arnold et al. 2005). Thus, the difficulty in

predicting a bacterial infection in febrile children could lead to the misuse of

antibiotics. Bacterial infections are associated with worsening of clinical conditions

with a consequent longer time for the resolution of symptoms. However, common colds

without secondary infections follow a self-limited course of seven-10 days with no

reduction of the duration of symptoms with the use of antibiotic, as it was observed in

this study and as reported elsewhere in the medical literature (Kaiser et al. 2001, Brandileone et

al. 2006, Li et al. 2009, Carranza-Martinez et al. 2010).

Although our study was not designed to evaluate the influence of fever in medical

decision to prescribe antibiotics, we observed that there was no difference in the

proportion of antibiotic prescriptions for patients who had fever at the onset of

symptoms and for those who did not, as well as in the proportion of judicious

prescription of antibiotics for these two groups of patients. Additional research is

needed to identify the local factors that are determinants of antibiotic misuse.

Although it is well known that influenza vaccination is the primary strategy to prevent

influenza, the Brazilian Health Ministry started public vaccination campaigns against

influenza for children in 2010 so that part of the population had no access to influenza

vaccine by the time of the study. None of the patients of the study were vaccinated

against influenza and this fact could explain the high frequency of influenza infection

in our patients (23.1%; 31/134), as showed in Table

I.

Antiviral treatment also plays an important role in decreasing influenza-related

morbidity and mortality, but none of 33 patients with laboratory confirmed influenza

infection received antiviral therapy, whereas all patients with influenza monoinfection

received antibiotics inappropriately (10/10) and those coinfected with influenza and

other viruses also received antibiotics inappropriately in a great proportion of cases

(71.4%; 5/7). A recent study in the United States of America reported that during

2012-2013, antiviral medications were underprescribed and antibiotics may have been

inappropriately prescribed to a large proportion of outpatients with influenza (Havers et al. 2014). The referred study emphasises

that the use of sensitive and specific tests for the rapid diagnosis of influenza and

other respiratory viruses is strongly recommended and could decrease antibiotic use and

guide appropriate use of antiviral agents in both outpatient and inpatient settings

(Bradley et al. 2011). In Brazil, diagnostic

tests for respiratory virus are still expensive and are not easily available for the

vast majority of the population, especially for low income people.

This study was conducted at one primary care facility and it is not possible to

generalise the results to all Brazilian primary health care services, however, these

preliminary data will help us to better understand antibiotic misuse among children with

common cold viral infections. We believe that continuing education on appropriate

antibiotics and antivirals use as well as accessibility to influenza vaccination,

sensitive and specific tests for the rapid diagnosis of respiratory viruses and

antiviral medication are essential to improve primary healthcare quality.