There is now overwhelming scientific evidence that central obesity, as opposed to total obesity assessed by body mass index (BMI), is associated with the most health risks and that the waist-to-height ratio (WHtR) is a simple proxy for this central fat distribution. This Opinion reviews the evidence for the use of WHtR to predict mortality and for its association with morbidity. A boundary value of WHtR of 0.5 has been proposed and become widely used. This translates into the simple screening message ‘Keep your waist to less than half your height’. Not only does this message appear to be suitable for all ethnic groups, it also works well with children.

Use of BMI
The Body Mass Index (BMI) has served us well as a proxy for obesity for many years, but it has always been recognised that it does not differentiate between the muscular and the overweight, except at very high BMIs. But there is an even more important problem with BMI. Even in the overweight, it is only a proxy for total fat in the body and it does not distinguish between individuals with different types of fat distribution.

The first BMI chart, which displayed BMI as a function of weight (horizontal axis) and height (vertical axis) using contour lines for different values of BMI, first appeared in 1981 in John Garrow’s book Treat Obesity Seriously[1]. Since the early 1980s, the classic BMI chart has been used extensively to assess the severity of obesity. Healthy weight for height is usually defined as a BMI between 18.5 and 25 kg/m2, overweight as equal to or more than 25 and less than 30, and obesity as a BMI of equal to or more than 30 [2].

Proposal to use waist-to-height ratio in primary screening to assess shape
The ratio (R) of the waist circumference (W)-to-height (Ht) (WHtR) was originally proposed more or less simultaneously in Japan [10] and the UK [11]-[13] as a way of assessing shape and monitoring risk reduction. Both proposers suggested that WHtR values above 0.5 should indicate increased health risk.

We believe that a simple index such as WHtR is a good proxy for central obesity and has great practical advantages. The greater propensity for South Asians to develop diabetes at lower BMI than white Europeans has been recognised for some time leading to different BMI ranges being suggested for South Asians [14]. The use of WHtR circumvents such problems because the adjustment of waist circumference for height means that the same boundary values are suitable for both ethnic groups.

Here we summarise the evidence that WHtR is a good predictor for morbidity and mortality and then discuss the practical aspects.

Simple practical screening with a shape chart based on WHtR
In the mid 1990s, one of us (MA) developed a chart (see Figure 1) based on WHtR (The Ashwell® Shape Chart) that allowed health professionals and/or their patients to match their waist circumference against their height- in inches or in centimetres- and to see into which category they fall [30]. Four regions based on boundary values for WHtR were designated. This is how they were described in terms of the ‘Action Steps’ for the patient:

• If your shape is in the ‘chilli’ region (WHtR less than 0.4), you should ‘Take Care’
• If your shape falls in the ‘pear’ region (WHtR between 0.4 and 0.5), you have a healthy ‘OK’ shape.
• If your shape falls in the ‘pear-apple’ region (WHtR between 0.5 and 0.6), you should ‘Consider Action’
• If your shape falls in the ‘apple’ region (WHtR above 0.6), your health is probably at risk. Why not talk with your health care provider, dietitian or practice nurse and ‘Take Action’?

Discussion
Complexity of using BMI and waist circumference to assess risk
One of us (MA) has argued before in favour of keeping screening methods simple [33]. A very good example of the complexity of setting cut-off values for waist circumference and BMI was apparent recently in a study which compared the relationship between adiposity and prevalence of diabetes across ethnic groups in the UK Biobank cohort [34]. The proposed ethnic-specific obesity cut-offs that equate to those developed on white populations in terms of diabetes prevalence are shown in Table 2. For men and women, the BMI values for different ethnic groups that are equivalent to BMI 30 in white men and women range from 21.5 to 26. The Action levels for waist circumference in men and women in the different ethnic groups are also smaller than their white counterparts and show great variability. We have added our proposed WHtR boundary value of 0.5 to this table to illustrate the universality and simplicity of this boundary value.

Simplicity of using WHtR to assess risk
There is enough evidence now from all the ethnic groups portrayed in Table 2 to suggest that WHtR 0.5 makes a perfectly acceptable global cut-off value. For example, this value has been used in recent papers from India [16],[17], in Korea [18], in China [19], in Sri Lanka [20], in Spain [21] and in Chile[22] and earlier studies from different ethnic groups were included in our meta-analysis [35].

In summary, we believe that a cut-off value of 0.5 for WHtR would be sufficient to indicate increased risk and that this value would be suitable for all ethnic groups. The inclusion of this simple value in Table 2 contrasts the simplicity of WHtR and the complexity of using cut-off values for BMI and waist circumference.

Cheap primary screening using a piece of string to assess WHtR
The adoption of WHtR 0.5 as the most important boundary value for simple primary screening allows freedom from sophisticated and expensive measuring devices. The advice to ‘Keep your waist to less than half your height’ means that a piece of string can be cut to represent a person’s height and the same piece of string can be used folded in half to see if it fits around that person’s waist. If it does, ok. If it does not, this simple screening method has shown that further investigations of cardiometabolic risk factors should be made.

Summary
The scientific evidence showing that WHtR is a better correlate of health risk than BMI is accumulating rapidly. These health risks include diabetes, hypertension, stroke, dyslipidemia and CVD. Translating science into policy always takes much longer. In this Opinion, we have presented new data to add to the case for using WHtR instead of BMI for primary screening purposes. We have shown that ten per cent of the whole population would be ‘missed’ if screening is only done on the basis of BMI. We have also pointed out the simplicity of measuring WHtR. If a tape measure is not available then a piece of string, which is folded so that it measures half a person’s height, can be used to show the ideal maximum waist circumference for health.

We make the plea that such a simple screening tool as a piece of string must be considered when the health risks from an obesity pandemic are acknowledged to be so great. Further, the problem is increasing rapidly in many countries which do not have access to more sophisticated measuring equipment.

Unwittingly, we have also come up with the answer to that age old question: How long is a piece of string? A piece of string could be cut to represent a person’s height and the same piece of string can be used (folded in half) to see if it fits around that person’s waist. If it does, ok. If it does not fit, then action to reduce the size of that person’s waist circumference is needed to reduce morbidity and mortality.

This article is adapted from Margaret Ashwell and Sigrid Gibson, A proposal for a primary screening tool: ‘Keep your waist circumference to less than half your height’. Source article. This work is licensed under a Creative Commons Attribution 4.0 License.