Risk factors for amputations in people with diabetes – a New Zealand study sheds new light

An enormous number of studies have investigated the risk factors for amputations in people with diabetes. Most risk factors found can be bundled into four groups, factors relating to: i) socio-demographic background (such as age or sex), ii) medical history (such as kidney disease or hyperglycaemia), iii) foot complications (such as peripheral arterial disease or previous amputation), and, iv) foot disease (such as foot ulcers or infection). Yet, no study had investigated combinations of these risk factors in all known people with diabetes in an entire nation over time.

That was until a new study published in the high-quality journal Diabetologia from our friends across the ditch – led by Dr Jason Gurney and colleagues from the Uni of Otago – followed all patients with diabetes in New Zealand for 3 years to see who had amputations or died, and what potentially caused them.

What they cleverly did was identify all ~220,000 people registered on New Zealand’s Virtual Diabetes Registry and linked them all to their hospital records for the previous 5 years. This allowed them to gain a comprehensive picture of each person’s socio-demographic, medical history and prior amputation profile. They then tracked their hospital records for another 3 years to see if they ended up having an amputation or died. This allowed them to control for many factors in statistical models to see what the independent risk factors for amputations were in a huge population with diabetes.

What they found were some already well-known risk factors and some lessor known ones. The well-known risk factors they found were: being male, Maori, poor, having chronic kidney disease, stroke, heart disease, peripheral vascular disease or previous amputation. The lesser known ones were: having chronic pulmonary disease, paraplegia or cancer. They attributed pulmonary disease to a history of smoking potentially resulting peripheral arterial disease, but did not attribute paraplegia or cancer. Nearly all these risk factors predicted both minor (below ankle) and major (above ankle) amputations; but, were stronger for major amputations. But that was just the start of what they found.

They also found the more co-morbidities a patient had the higher the risk they were of having an amputation. To find this out they used a well-known measure of a person’s co-morbidity profile that predicts death at 10 years; the Charlson comborbidity index. This index shows that the more co-morbidities (from 17 major medical conditions) you have the more likely you are of dying sooner. They used this index to see if it also predicted amputations and it did just that.

Put simply this indicates the sicker a person with diabetes is the more likely they are of having an amputation. This suggests that perhaps we can use this index to not only predict early death, but also to predict early “limb death” (i.e. amputation). And speaking of death, they also found, like many others have, very high mortality rates in those having amputations; 57% of those having a major amputation and 35% of those having a minor amputation died within 3 years.

The authors concluded, “Using a large, well-defined, national cohort of people with diabetes, we found that being male, indigenous Māori, living in deprivation, having a high comorbidity burden and/or having a previous amputation were strongly associated with subsequent risk of lower limb amputation.” While a “previous amputation … may have less to do with the prior amputation per se than it does with the individual’s underlying risk factors”.

This study had many strengths – it was able to identify all people with diabetes in a whole nation, track those people’s hospital records, amputation procedures and deaths for many years before and after any amputations and control for a whole range of potential risk factors – however, it did have limitations. These included: they could not differentiate type 1 and type 2 diabetes, they didn’t measure peripheral neuropathy or foot disease, peripheral vascular disease combined arterial and venous disease, and they did not have access to primary care records which may have given a more accurate profile of patient’s medical history and foot disease history and importantly the care they had received.

Regardless, these findings shed more robust light on the risk factors for amputation, and perhaps we are getting closer to developing prognostic tools for our patients with foot disease – just like the Charlson comborbidity index – to predict recurrence, amputation and death; similar to what happens with cancer patients today. Imagine a future where you could tell your patient “from the foot disease calculator we estimate you have a 45% chance of having a major amputation in 3-years if we do nothing, but a 15% chance if you undertake intensive foot care at an accredited foot disease unit”.

What this paper beautifully reinforces is the need to use the data we already have much smarter to better help our patients and their clinicians on their pathway to avoiding amputations as advocated in the Australian diabetes-related foot disease strategy. Data that can help us develop future accurate prognostic tools to predict avoidable amputations in Australia can only bring us a step closer to our national goal of ending avoidable amputations in a generation … watch this space.