Navigating uncertainty: managing respiratory tract infections

---

Key practice points:

• Most RTIs are of viral origin, rather than bacterial, and in either case, the infection is likely to be self-limiting and the negligible benefit of antibiotic treatment does not outweigh the potential harms
• Provide patients or caregivers with information about the expected duration of RTI symptoms, advice for symptomatic relief and when to seek further medical advice
• Antibiotics may be considered for people with a suspected or confirmed bacterial RTI who are at high risk of complications, e.g. those with immunosuppression, or in specific clinical scenarios, e.g. all cases of bacterial pneumonia, bilateral acute otitis media in children aged under two years
• Always consider the possibility of meningococcal disease in people presenting with influenza-like illness; ensure that patients and/or caregivers know when to seek urgent medical advice
• Throat swabs and antibiotic treatment for Group A Streptococcus infection are only indicated for people who present with sore throat who have risk factors for rheumatic fever

Uncertainty in general practice: a familiar foe

“If there is one certainty in general practice, it is the inherent presence of uncertainty.” ¹

Diagnostic uncertainty is an unavoidable aspect of medical practice. Non-specific presentations, unclear or conflicting examination or test results, and inconsistent treatment responses are not uncommon.¹ The way in which uncertainty is navigated can significantly impact prescribing practices, the ordering of tests and other clinical investigations, and patient interactions and expectations.

Following through a set strategy can help to manage uncertainty, for example:¹

• Establish what symptoms the patient is most concerned about and what outcome they are seeking
• After history and examination, rule out red flags/serious diagnoses
• Consider the need (if any) for investigations, use clinical risk prediction tools, consider the “test of time” for a diagnosis to emerge
• Put a plan in place for follow-up and review if symptoms deteriorate or do not resolve

Uncertainty may contribute to inappropriate antibiotic prescribing

Every winter in New Zealand there is an increase in the number of people presenting to primary care with symptoms such as cough, sore throat and ears, and sinus congestion, usually as the result of a respiratory tract infection (RTI) (See: “Why do we get sicker in winter”).^2,3 Antibiotic dispensing is 30–40% higher in winter than summer in New Zealand; this is more than double the seasonal variation in the United Kingdom, suggesting some people are receiving antibiotics unnecessarily.^2,3

Most RTIs are caused by viruses and are self-limiting, and antibiotics are not appropriate. However, in some cases it is difficult to be certain if an infection is viral or bacterial, who would most benefit from antibiotic treatment and the likelihood of complications.⁴ Concern about missing a significant infection which does require antibiotics, e.g. bacterial pneumonia, can influence the prescribing decision, along with other factors such as patient expectations about receiving an antibiotic, or clinicians’ assumptions about these expectations, and the ability or likelihood of a patient to return for a follow-up appointment.

Antibiotics are usually unnecessary for “winter ills”

There are two main decision points when determining whether antibiotic treatment is indicated for a patient presenting with an RTI:

1. Is the infection viral or bacterial?
2. If bacterial, is antibiotic treatment indicated (and which antibiotic and for how long) and is it needed for this particular patient?

In a typical primary care setting, it is often not possible to definitively determine the type of infection (viral versus bacterial). Instead, clinicians rely on their clinical findings and judgement to make a presumption about the infection type and consequently whether antibiotic treatment is indicated.

Bacterial infection may be suspected in patients who are very unwell with persistent and/or worsening symptoms (see Table 1 for details), however, this can be difficult to establish in patients who present early in the course of illness. Investigations such as requesting a full blood count and C-reactive protein (CRP) level may assist with diagnosis in some circumstances (see: “Using CRP testing to guide antibiotic prescribing”). Even if an infection is presumed to be bacterial, and antibiotics are indicated for that infection, not all patients will require antibiotics, e.g. otitis media or sore throat in people without risk factors. Most RTIs are self-limiting and the benefits of treatment will often not outweigh the risks, e.g. adverse effects such as nausea, diarrhoea, rash, Candida albicans or Clostridium difficile infection, hypersensitivity reactions and development of antibiotic resistance, both in the individual and population. For example, a Cochrane review on antibiotic use in children for otitis media found that the number needed to treat for a small reduction in pain at two to three days was 24, but the number needed to harm (defined as vomiting, diarrhoea or rash) was 13.⁶ Overall, antibiotics made little difference to the outcomes of children with acute otitis media in terms of effects on pain, short-term hearing loss or eardrum perforations.⁶

For further information on the limited effectiveness of antibiotics for winter illnesses, see: www.bpac.org.nz/2018/cold-season.aspx

N.B. Clinicians should have a high level of suspicion of meningococcal disease, i.e. meningococcal meningitis and meningococcal septicaemia, in people presenting with influenza-like illness (see: “Consider meningococcal disease in people presenting with non-specific influenza-like symptoms”).

Communicating the decision not to prescribe antibiotics

Patients or caregivers presenting to primary care are generally seeking symptomatic relief and reassurance that they, or their child, do not have a more serious illness. Asking “what can I do for you today” is a useful conversation opener because it allows the patient to describe the symptoms they are most concerned about and to express their expectations for treatment. Although there has been much work done on increasing understanding about the role of antibiotics, there are still many patients that believe that antibiotic treatment is necessary for a RTI, and they may express frustration or disappointment that a prescription is not offered.

Effective communication is the most important treatment for uncomplicated RTIs. Ensure the patient feels validated for presenting for assessment and exp lain that just because an antibiotic is not indicated, it does not mean they are not unwell.

Discussion should cover:

• What illness they are likely to have
• The expected duration of symptoms (Table 1); a medical certificate may be required
• The role of antibiotics in RTI; most infections are viral and self-limiting, antibiotics will not help their symptoms and may cause harm, e.g. adverse effects, antibiotic resistance
• Symptomatic management advice (Table 1); a prescription for a supportive treatment may be appropriate
• A plan for review if symptoms persist or worsen, e.g. a follow-up appointment or a phone call to the practice
• When to seek urgent medical care, e.g. shortness of breath, stridor, fever >40°C, neck stiffness and rash

Patient information is available here:

• “A cold or the flu?” – practical information and links to other resources: www.healthnavigator.org.nz/health-a-z/c/cold-or-the-flu/
• “Keep antibiotics working” – explaining antibiotic resistance: www.pharmac.govt.nz/keepantibioticsworking
• “What to do when you have a cold” – instructions for symptomatic relief and infection prevention: www.goodfellowunit.org/sites/default/files/webinars/common_cold_leaflet_2017.pdf

When should antibiotics be considered?

Specific scenarios where antibiotic treatment is indicated or should be considered for people presenting with a suspected or confirmed bacterial RTI or associated condition are outlined in Table 1.

Ideally, the decision to prescribe or not prescribe antibiotics should be made based upon the patient’s current presentation, rather than providing a prescription to fill later, e.g. if the patient’s condition deteriorates or they remain unwell (often called a “back pocket” prescription). While this practice is sometimes recommended as a strategy to reduce antibiotic prescribing, there may still be patients who receive antibiotic treatment unnecessarily, e.g. if they fill the prescription immediately without waiting to see if their symptoms improve. This practice also transfers decision-making responsibility to the patient when it should be the clinician’s decision whether antibiotic treatment is indicated. However, there may be exceptions and clinical discretion applies, e.g. if there are socioeconomic factors which influence the ability of the patient to attend a follow-up appointment. The decision and reasoning should be clearly documented in the patient’s notes.

The threshold for considering antibiotics for patients with a suspected or confirmed bacterial RTI should be low for those who are at high risk of complications due to compromised immune function, such as people with:⁸

• Immune suppression
• Significant heart, lung, renal, liver, metabolic, autoimmune or neuromuscular disease, e.g. heart failure, COPD, bronchiectasis, severe asthma, diabetes
• History of premature birth (younger children)
• Frailty
• History of hospitalisations, particularly in the past year

N.B. Ensure that the same threshold for prescribing is applied to PSO antibiotics as for a regular prescription.

For information on antibiotic prescribing regimens for respiratory tract and ear infections see: www.bpac.org.nz/antibiotics/guide.aspx

For further information on symptomatic treatments, see: www.bpac.org.nz/2018/cold-season.aspx

---

Table 1. Upper respiratory tract infection (URTI) symptoms and associated conditions, duration of symptoms, indication for antibiotic treatment and supportive treatments⁹

^† May improve symptoms but is unlikely to influence symptom duration, and could cause systemic adverse effects; patients should be advised to only use the prescribed dose.

^‡ A Cochrane review of the evidence for antibiotic treatment of acute otitis media found that pain (primary outcome) recovered spontaneously in 60% of children within 24 hours, whether they received placebo or antibiotics. Antibiotic treatment resulted in a 5% absolute risk reduction in pain at two or three days; 6% absolute risk reduction in pain at four to seven days 6

Using CRP testing to guide antibiotic prescribing

Serum C-reactive protein (CRP) is an acute phase marker for a number of pathological processes, including infection and inflammation. Healthy people usually have a CRP of < 5 mg/L; CRP increases within four to six hours following infection and peaks at around 36 hours. The level of elevation usually corresponds to the severity of the infection. Although the test cannot definitively distinguish the type of infection, bacterial infections generally cause a greater elevation in CRP than viral infections.

For example, CRP levels can be used to help clinicians decide whether or not to prescribe an antibiotic to patients where a diagnosis of pneumonia is uncertain, e.g. the patient presents with cough and one or more other associated symptoms such as fever, sputum production, breathlessness, wheeze, chest discomfort or pain.¹⁹ N.B. chest x-ray is the gold standard for confirming pneumonia, but is not necessary for a diagnosis

If a point-of-care testing device is available, a CRP level may also be useful to support advice to the patient that no antibiotic treatment is required, e.g. patients with symptoms of self-limiting upper RTIs.

The United Kingdom National Institute for Health and Care Excellence has made the following recommendations about CRP testing to guide antibiotic prescribing:¹⁹

• CRP < 20 mg/L suggests a viral or mild bacterial infection; antibiotics should not be offered routinely
• CRP 20–100 mg/L can be difficult to interpret; consider antibiotics if the clinical findings (history and examination) suggest a bacterial infection or the patient is at high risk of complications, however, a CRP level in this range is also not unusual for someone with influenza.
• CRP > 100 mg/L is indicative of a significant bacterial infection and antibiotics should be recommended

For further information on point-of-care CRP testing, see: www.bpac.org.nz/bpj/2015/june/crp.aspx