A total of 71,000 individuals presenting with a DFU were included in the study. These patients had a median age of 71 years (IQR 60–80); 59,425 (83.7%) had type 2 diabetes and 9520 (13.4%) had type 1 diabetes; the remaining 2050 individuals (2.9%) had other or unknown types of diabetes (Table 1). The median duration of diagnosed diabetes was 15 years (IQR 9–20), with 1945 (2.7%) having a missing date of diagnosis. Ethnicity data were missing for 6790 individuals (9.6%); 58,275 individuals (82.1%) were from White ethnic groups, 3120 (4.4%) were from Asian ethnic groups, 1630 (2.3%) were from Black ethnic groups, and 1190 (1.7%) were from mixed and other ethnic groups. Individuals were not evenly spread across the spectrum of socioeconomic deprivation: 17,250 (24.3%) were living in the most deprived quintile and 9855 (13.9%) in the least deprived quintile; data were missing for 1855 (2.6%).

Across the whole cohort, the median SINBAD score was 2 (IQR 1–3), with 30,605 (43.1%) having a score of 3 or more. Within the SINBAD sub-categories, 34,500 (48.6%) had an ulcer greater than 1 cm2 in area, 28,125 (39.6%) had evidence of a bacterial infection, 11,995 (16.9%) had a deep ulcer (to muscle/bone), 13,680 (19.3%) had an ulcer on the hindfoot, 24,795 (34.9%) had evidence of peripheral arterial disease (ischaemia), and 53,440 (75.3%) had recorded evidence of neuropathy.

Of all the individuals included in the analysis, 1450 (2.0%) had one or more hospital admissions for MI in the year preceding first registration of the index DFU by the specialist footcare team. The corresponding numbers for heart failure and stroke were 8250 (11.6%) and 1775 (2.5%), respectively. A total of 2385 individuals (3.4%) had received RRT in the year preceding first registration, and 17,650 (24.9%) had evidence of CKD at stage 3 or greater but had not received RRT.

When compared to all people with diabetes included in the 2019/2020 NDA, people in the cohort presenting with a DFU were more likely to be male (67.7% vs 55.7%, p<0.0001), older (median age 71 years vs 65 years, p<0.0001), and to have a longer duration of diagnosed diabetes (median 15 years vs 8 years, p<0.0001). Those with a DFU were also more likely to be of White ethnicity (82.1% vs 66.3%, p<0.0001). Although there were statistically significant differences in the distribution across deprivation quintiles in the people with a DFU compared with all people in the 2019/2020 NDA (p<0.0001), the differences were not substantial (24.3% of people with a DFU live in the most deprived quintile of areas compared with 23.8% of all people in the 2019/2020 NDA).

Within the cohort of 71,000 people registered with a DFU, 2985 died within 12 weeks (4.2%; 95% CI 4.1, 4.4), 5850 died within 26 weeks (8.2%; 95% CI 8.0, 8.4), and 10,210 died within 52 weeks (14.4%; 95% CI 14.1, 14.6). In those aged 80 years and older at the FEA, the mortality rates at 12, 26 and 52 weeks were 9.1% (95% CI 8.7, 9.5), 17.0% (95% CI 16.5, 17.5) and 28.4% (95% CI 27.7, 29.0), respectively.

Across the whole cohort presenting with DFUs, the expected 12-week and 26-week mortality rate exceeded 25% for 1185 individuals (1.7%) and 5165 individuals (7.3%), respectively. In those aged 65 years and older at the FEA, 2.5% had an expected mortaltiy rate of 25% or greater at 12 weeks and 11.5% had an expected mortality rate of 25% or greater at 26 weeks. In those aged 80 years and older at the FEA, the proportions with an expected mortality rate of 25% or greater were 5.5% at 12 weeks and 22.1% at 26 weeks.

Factors associated with mortality were similar at 12 and 26 weeks (Figs 1 and 2). Older age was the variable most closely associated with higher mortality rates. At 26 weeks, individuals aged 80 years and older had a mortality rate ratio of 2.15 (95% CI 2.03, 2.28) compared with those aged 65–79 years. There were no statistically significant differences in mortality rates by sex, socioeconomic deprivation or duration of diagnosed diabetes at either 12 or 26 weeks. People with type 1 diabetes had lower mortality rates (rate ratio 0.90; 95% CI 0.81, 1.00) than those with type 2 diabetes. People from Black ethnic groups had lower mortality rates than those from White ethnic groups (rate ratio 0.80; 95% CI 0.66, 0.96). Ulcers with an area of 1 cm2 or greater (rate ratio 1.50; 95% CI 1.42, 1.59), deep ulcers (rate ratio 1.26; 95% CI 1.18, 1.35), ulcers on the hindfoot (rate ratio 1.53; 95% CI 1.44, 1.62) and evidence of peripheral arterial disease (ischaemia) (rate ratio 1.78; 95% CI 1.69, 1.88) were all associated with a higher mortality rate at 26 weeks. Recorded evidence of bacterial infection and neuropathy were not statistically significantly associated with mortality rates.

Rate ratios (95% CI) for mortality within 12 weeks of first registration of a DFU with a specialist footcare team

Rate ratios (95% CI) for mortality within 26 weeks of first registration of a DFU with a specialist footcare team

At 52 weeks, older age was also associated with higher mortality rates (rate ratio 2.11; 95% CI 2.01, 2.20 for those aged 80 years and older compared with those aged 65 to 79 years old), but, as at 12 and 26 weeks, there was no statistically significant association with sex (Fig. 3). Type 1 diabetes was associated with a lower mortality rate than type 2 diabetes (rate ratio 0.92; 95% CI 0.86, 1.00); the association between longer duration of diagnosed diabetes and higher socioeconomic deprivation with higher mortality rates reached statistical significance (p<0.05 and p<0.05, respectively); people from Black and Asian ethnic groups had a lower 52-week mortality rate than White ethnic groups (rate ratio 0.86; 95% CI 0.74, 0.98 and rate ratio 0.87; 95% CI 0.78, 0.97, respectively) and the associations between mortality and ulcer characteristics and cardio-renal comorbidities were similar to those at 12 and 26 weeks.

Rate ratios (95% CI) for mortality within 52 weeks of first registration of a DFU with a specialist footcare team

As an illustration of the divergent outcomes predicted by the model, Table 2 shows the expected mortality rates for four hypothetical people. These hypothetical people are presented as examples but are based on people with similar characteristics seen by the authors.

Patient 1: A man aged 40–59 years from an Asian ethnic group and in the second most deprived socioeconomic quintile who has been diagnosed with type 2 diabetes for 5–9 years and who presents with a foot ulcer with the highest SINBAD score, with no hospitalisation for MI, heart failure, stroke or RRT in the previous year but CKD at stage 3 or greater, has an estimated 12-week mortality rate of 3.3% and a 26-week mortality rate of 7.0%.

Patient 2: A White man aged 65–79 years in the third most deprived socioeconomic quintile who has been diagnosed with type 2 diabetes for 20–29 years who presents with a non-infected forefoot foot ulcer that is over 1 cm2 in area and extends to the muscle or tendon, with evidence of ischaemia and neuropathy, a history of hospitalisation for heart failure and stroke in the previous year and CKD 3 or higher, is estimated to have a 12-week mortality rate of 17.6% and a 26-week mortality rate of 34.9%.

Patient 3: A woman aged 80 years or older from a White ethnic group living in one of the most socioeconomically deprived areas, who has been diagnosed with type 2 diabetes for between 10 and 19 years, who has a foot ulcer smaller than 1 cm2 on the hindfoot that does not extend to the muscle and tendon, with evidence of ischaemia and neuropathy, with hospital admission for MI and stroke in the previous year and CKD at stage 3 or greater, is estimated to have a 12-week mortality rate of 16.9% and a 26-week mortality rate of 37.5%.

Patient 4: A White man aged 80 years or older in the second most deprived socioeconomic quintile with type 2 diabetes of 10–19 years duration, the most severe foot ulcer (maximum SINBAD score), no hospitalisation for cardiovascular disease but receiving RRT in the previous year, has a 12-week mortality rate of 29.8% and a 26-week mortality rate of 65.2%.