Respiratory Infections and Antibiotic Decision-making
By Dr. Christopher Cirino
Key points: Antibiotics are often used in the outpatient clinic for viral infections. Antibiotic decision-making is subject to a form of bias in both physician and patient, confirmation bias, and makes opaque an already difficult set of decisions. This article will review subjective findings that sometimes are rooted in this bias and how a patient can partner with their doctor in ensuring that antibiotic therapy is justified.
As an infectious diseases physician, this is a topic near and dear to me. Most patients presenting to the clinic, ER or urgent care with “cough” have a viral syndrome or post-viral bronchitis/reactive airways disease. Antibiotics are being given – even to these patients who have illnesses that usually improve with a tincture of time (and chicken noodles soup) and that antibiotics will not improve. These medications also have risks, including side effects and more severe consequences.
Behind these practices lies a behavior that has been reinforced both in the clinician prescribing and the patient requesting — thus creating an altered perception, or bias. This post will outline the basis of this problem, summarize typical signs and symptoms of these infections and advise how a patient can protect themselves and ensure appropriate care.
A clinician will have seen thousands of patients with viral syndromes by the time they complete formal training It is the most common cause of cough with no localized pulmonary signs/symptoms. Even for patients presenting with a pneumonia as indicated on a radiograph, virus were still more readily detected than bacteria. In one study on the cause of pneumonia in 2,300 adults, 27% were from a virus (most commonly rhinovirus and then influenza in this study) and 14% from a bacteria (most commonly Streptococcus pneumoniae). That means that almost 62% of patients had no determinable (to the routine tests) pathogen. Likely the majority of these were viral causes, though some could be caused by more difficult to grow bacteria. Rapid molecular testing has decreased this unknown percentage and will likely continue to overtake more traditional laboratory methods.
Most patients who come in the clinic often have to schedule their visit a few days earlier and generally are not as acutely ill as someone having to go to the ER. They may have a nagging cough that has persisted, ongoing drainage. The majority of these patient visits for respiratory tract infections are for viral infections. Clinicians in a busy clinic have the task of triage – deciding whether a patient’s history and physical findings require further testing, e.g. a radiograph or bloodwork, or antibiotics.
Bias affects decision making when there are multiple variables, time constraints and expectations. Doctors occasionally rely on a syndromic characterization — quickly determining if the process warrants further work-up. The experience-base of the clinician, since it is drawn upon by specific patient presentations, is unfortunately a glimpse of a disease manifestation as it relates to a few patients — and may be a framework of bias. Clinicians may listen for key words such as “green sputum”, “sweats”, “fever” and begin to consider whether antibiotics may be needed. Clinicians may confirm that a patient improved after antibiotics in the past when they receive a call “it made me better by the next day”, not factoring in other characteristics, such as the coinciding improvement of a viral infection or anti-inflammatory properties of some antibiotics (e.g. azithromycin, levofloxacin,etc) or from other things taken or prescribed. This type of bias is known as Confirmation bias. Not seeing alternative stories as valid and fixing on cause and effect anticipated is confirmation bias.
This form of bias does not only come from the clinician but also from the patient — sometimes even before they enter the clinic! I frequently see patients who are concerned, even convinced, that their respiratory infection requires an antibiotic. Occasionally, they will call the office, requesting that an antibiotic be ordered, as if they are ordering take-away from a fast food restaurant. Although it is not my objective to hold therapy when therapy is needed, I find that the overwhelming majority of patients that call or come into my office for respiratory infections do not need an antibiotic. They will often state that in the past, when they received an antibiotic — or that they almost waited too long and they were having a worsening cough — that the antibiotic worked. Not only are they equating their current illness with what happened in the past but they are looking for characteristics that are not specific enough to discern between a viral and bacterial infection.
How do these misconceptions arise?
These beliefs are intrinsically connected with an individual’s prior encounters with the medical system, lack of knowledge of what differentiates a viral from a bacterial infection, and perhaps physician-patient enculturation in the “antibiotic era”.
If a patient had a respiratory illness and received an antibiotic in the past, they will attribute the spontaneous, gradual resolution of the viral infection to the potent antibiotic regimen that was given to them. These past positive experiences (post-antibiotic yeast infections forgotten or forgiven) fuel future expectations. It is why patients occasionally use the statements “Well I get something like this every year, but only when the doctor writes me a Z-pack, it gets better.”
How common does a doctor write an antibiotic prescription for a respiratory infection?
In one study of 366 pediatric and family physicians in Georgia polled, 38% (pediatrics) and 58% (family practice) indicated routine prescribing for the common cold; 81% and 93% routinely prescribed antibiotics for “bronchitis”. In a chart review of this same study, 19% of antibiotic prescriptions were called in after telephone encounters.
Perceived or real pressures placed on physicians by parents or patients themselves was cited by more than 50% of the physicians polled.
Factors that influence antibiotic decision-making or seeking which are not specific enough:
The color of sputum/mucus:
This is a common concern that is brought up by patients and sometimes a reason why clinicians order antibiotics. Most patients would not seek physician consult if their secretions were clear. However, the shift from clear to yellow- or green-colored secretions does not increase the probability that a bacterial infection exists, except possibly in patients with chronic obstructive lung disease (COPD). Though it leads to greater patient preoccupation.
The yellow/green color of mucus originates from inflammation and is directly related to leukocytes (white blood cells) migrating to the area of infection.
2. The duration of the cold symptoms:
The length of time of the cough or cold symptoms (if cold symptoms think virus) has little bearing in differentiating a viral infection from a bacterial infection. Viral infections and their aftermath may cause cough and drainage symptoms for several weeks, particularly if one develops a reactive airway disease (aka bronchitis). Occasionally, one may even be co-infected with another viral infection which may be a factor of severity. I would be more concerned if a patient develops sudden onset of fever with rust-colored sputum and chest pain after improving from a viral syndrome – this is the classic presentation for Pneumococcal or Staph aureus pneumonia. Otherwise, setting an arbitrary time-point in deciding on antibiotic therapy is a common misconception by both clinician and patient.
3. Underlying diseases such as COPD or mild immunosuppression with a viral infection:
Although chronic lung disease increases concern and risk of respiratory issues, these conditions do not in and of themselves merit consideration for antibiotic therapy when the patient only has a viral infection. One exception is a more significant exacerbation of COPD: studies have supported the role of azithromycin in exacerbations (possibly from the anti-inflammatory role). Even viral illness can cause enough issues to lead to hospitalization in a patient with more severe disease. Treatment may require stabilizing the reactive airways, including a COPD exacerbation, with bronchodilators (albuterol) and inhaled or oral steroids or a brief course of IV steroids. The goal in these patients is to rule out concomitant bacterial infection often with a radiograph if their clinical examination alone is challenging to sort out and alert the patients for warning signs to seek further care – and they come in when they are worse — and get treated.
What are the Clinical exam factors between viral and bacterial infections?
Discerning between a viral infection and a bacterial infection is usually possible with an accurate history and physical. A patient with a viral process will present with a runny nose, sore throat, ear symptoms, muscle aches and cough. A bacterial process may develop acutely with chest pain (not just from coughing as in bronchitis) possibly one-sided, less commonly bilateral, persistent fever and productive rust-colored sputum. The most common history for bacterial pneumonia is “double hump” fever or presentation — where a patient had a viral syndrome and improving and then develops a bacterial pneumonia. The virus is weakening the tissue of the respiratory tree and contributing to bacterial overgrowth and invasion (usually of normal flora = Strep pneumoniae or Staph aureus, less common Klebsiella). Less commonly is a bacterial superinfection in the setting of findings of a viral infections – and is more commonly seen with influenza. The least common is a bacterial infection acquired from the environment e.g. Legionella, Chlamydophila psittaci, Coxiela, etc) or an acute fungal pneumonia (Cryptococcus, Coccidiodes,etc)
In both a bacterial infection or viral syndrome, a patient may have fever and chills. There may be an elevated pulse and respiratory rate in more serious viral pneumonitis or bacterial pneumonia.
The examination often reveals nasal membrane swelling and mucus bridging in the nares. The posterior pharynx (back of the throat) is often inflamed, with some glandular tissue swelling or “cobblestoning”. Exudates or white patches within the pharynx may be seen in viral infections (mononucleosis, adenovirus, etc) and are not confined to Strep throat alone. Occasionally, there will be some mucosal drainage in the posterior pharnx. The neck will often have some lymphadenopathy or lyphadenitis. (enlarged, tender lymph nodes).
The lung exam will occasionally reveal wheezing. The auscultation of rales (or wet, crackling sounds) raises concern for a bacterial infection, though occasionally viral infections (eg. flu, adenovirus) can cause a significant, even life threatening pneumonitis. I will often request a chest radiograph upon hearing these sounds. If nothing is seen and pneumonia is not confirmed (CXR will not pick up 20% of those with bacterial pneumonia at outset, then I have to weight options and decide to treat or not to treat. Either way, close follow-up is paramount.
What are The Harms of Antibiotic exposure?
Yeast infection owing to a shift in skin flora to yeast predominance and infection.
Clostridium difficile, a potentially serious overgrowth process where a toxin is secreted by the bacteria and causes diarrhea and sometimes serious bowel swelling or colitis.
Antibiotic toxicity, not only nausea, vomiting, rash, but also a severe, exfoliating rash (Steven-Johnson’s Syndrome)
Shift/Selection of skin and mucosal flora that are resistant to antibiotic; a close link has been found in patients receiving antibiotics and the development of Methicillin Resistant Staphylococcus aureus. I see the association with MRSA colonization after cephalosporin or fluoroquinolone use.
Altered gut flora and association with obesity. There has been an association found with the use of antibiotics during childhood and risk for obesity.
What can you do to prevent the inappropriate use?
Avoid seeking care if symptoms are consistent with a viral infection. If many people are sick in the household, it is exceedingly unlikely that it is a bacterial infection. That includes even in circumstances, when a physician decided that one of your children has an ear infection and gave him/her antibiotics (though I would question whether that child REALLY has a bacterial infection).
Ask the physician whether it is necessary for antibiotics if you are found to have a cold or a bronchitis.
Arrange an appointment if the cough is persistent or if you have developed bronchitis/asthma exacerbation. Usually a bronchodilator like albuterol or a short course of an inhaled steroid/bronchodilator mixture is enough to improve the condition over the course of 5-7 days.
By all means, if you have persistent fever, shortness of breath or chest pain, you need to be seen! A lingering post-viral cough that suddenly worsens to include chest or back pain, high fever and productive sputum makes bacterial infection more likely. Usually the process comes after the lung tissue is injured by the virus. Bacteria are capable of growing and are not mobilized out because of the injured tissue and then cause an acute pneumonia.
In summary, bias can affect antibiotic decision-making and requests. Most outpatients that are not sick enough to go to the ER are presenting with viral syndromes. Patients need to ensure that they are getting appropriate treatment by being actively involved in the decision process.
Have an uneventful flu season. Thanks for reading this. Please share this!