Making Sense of Lung Function Tests

Jonathan Dakin, Mark Mottershaw, Elena Kourteli

• Editorial: CRC Press
• Año: 2017
• ISBN:
  9781315355351
• Edición: 2
• Categoría: Medicina
• Categoría: Medicina polmonare e toracica

Respiratory problems are the most common cause of acute admission to hospital. A variety of diagnostic investigations are required, both for acute and clinic assessment. Making Sense of Lung Function Tests, Second Edition familiarises both trainees and more experienced clinicians with the interpretation of a range of respiratory parameters. It places lung function in a clinical context using real-life examples and provides invaluable hands-on guidance.

For this second edition Consultant Respiratory Physician Jonathan Dakin and Consultant Anaesthetist Elena Kourteli are joined by Mark Mottershaw, Chief Respiratory Physiologist from Queen Alexandra Hospital, Portsmouth, all contributing a broad range of expertise and perspectives. Together they have updated the book throughout and added new chapters including an algorithm for interpretation of pulmonary function tests, exhaled nitric oxide (FENO) and cardiopulmonary exercise testing.

The text offers a clear explanation of the concepts which students find difficult, including:

• 
  
  

• The basis of obstructive and restrictive defects

• 
  
  

• Pattern recognition of the flow volume loop

• 
  
  

• Differences between TLCO and KCO

• 
  
  

• Assessment of oxygenation using PO2 and SO2

• 
  
  

• The basis of Type 1 and type 2 respiratory failure

• 
  
  

• Distinguishing respiratory and metabolic acidosis

• 
  
  

• The relationship between sleep and respiratory failure

The information is presented in an accessible way, suitable for those seeking a basic grounding in spirometry or blood gases, but also sufficiently comprehensive for readers completing specialist training in general or respiratory medicine.

1

---

---

Expressions of normality

The percentage predicted has long been the favoured method of expressing lung function results amongst clinicians. It has the advantages of being easy to calculate and intuitive to understand. A test result which falls below 80% of the predicted value is often considered to be outside the range of natural variability and therefore abnormal, for a number of pulmonary function indices.

The percentage predicted is also used to grade severity of disease by comparing test results with a table of cut-off ranges. The number of categories and exact cut-offs are fairly arbitrary and vary between different respiratory societies. For example, one such table for identifying abnormal spirometry based on the FEV₁% predicted is shown in Table 1.1, modified from the American Thoracic Society (ATS)/European Respiratory Society (ERS) taskforce guidelines on interpretative strategies for lung function testing.¹ A similar classification is in common usage for peak flow readings in asthma (Table 2.1).

However, different lung function tests and indices have different degrees of natural variation within the population. For example, the transfer factor for carbon monoxide (TL_CO) has a wider inter-individual variability than many other lung function test values, and therefore a result which is 75% predicted may be well within the normal range. Moreover, this normal range may alter with age, so a value which is 75% of that predicted may be normal in the elderly, but warrant further investigation in the young.

This shortcoming has led clinical physiologists to favour the concept of the standard residual as a statistically more valid approach to identifying normal ranges. This method involves using standard deviations (SDs) to identify the upper and lower limits of normality (ULN and LLN respectively). Figure 1.1 shows a typical bell-shaped normal distribution curve and includes the percentage of values which lie within each SD (or Z score) and the mean. In a normal distribution, 95% of the population will record values within two SDs above or below the mean value.

The convention amongst physiologists is to use a value of 1.64 SDs to identify the ULN and LLN. This value is chosen because in a normal distribution 90% of the population will fall within ±1.64 SDs of the mean, with 5% having ‘supranormal’ values above this range and 5% having ‘abnormal’ results below this range. However, there is no pathology associated with a supranormal value (with few exceptions such as measurements of airway resistances – see Chapter 8) and those fortunate individuals may be placed within the normal range, from a medical point of view. Therefore, the limit of –1.64 SDs below the mean identifies a 95% confidence limit, below which measurements are abnormal. The standard residual may be used to express the distance a result lies from the mean, and thereby grade the severity of abnormality, as shown in Table 1.2.

Table 1.1 Severity of airflow obstruction by FEV₁

Degree of severity	FEV1% predicted
Normal	>80
Mild	70–79
Moderate	60–69
Moderately severe	50–59
Severe	35–49
Very severe	<35

Figure 1.1 Normal distribution curve showing the percentage of a normal population who would fall 1.64 standard residuals beneath the mean. If the limit of normality is placed at 1.64 SDs below the mean, the healthy range encompasses 95% of the population.

Table 1.2 Grade of severity by standard residual

Standard residual	Grade of severity
–1.64 or greater	Normal
–1.65 to –2.50	Mild
–2.50 to –3.50	Moderate
<–3.50	Severe

 KEY POINTS

• The percentage of predicted is the most commonly used expression of normality, which is simple to calculate and intuitively understood. However, the cut-off for normality (e.g. <80%) is chosen arbitrarily and may result in under- or overdiagnosis of pathology.
• The use of standard residuals is more robust and provides a statistically valid method to identify values that fall below the limits of normal physiological variability. Usage of standard residuals is increasing and may ultimately replace the percentage predicted.

PART

1

TESTS OF AIRWAY FUNCTION AND MECHANICAL PROPERTIES

2 Peak expiratory flow

3 Spirometry and the flow–volume loop

...