]ourrial of Internal Medicine 199 3 : 2 3 3 : 48 5-49 1

Long-term prognosis for diabetic patients with foot ulcers

J. APELQVIST, J. LARSSON* & C.-D. AGARDH From the Lkpartments of Internal Medicine and *Orthopdic Surgery, University Hospital, Lund. Sweden

Abstract. Apelqvist J , Larsson J. Agardh C-D (The Departments of Internal Medicine and Orthopaedic Surgery, University Hospital, Lund, Sweden). Long-term prognosis for diabetic patients with foot ulcers. Iournal of Internal Medicine I 99 3 : 233 : 4 8 5-49 1 .

Objective. To evaluate the recurrence of foot ulcers as well as the cumulative amputation and mortality rates in diabetic patients with previous foot ulcers. Design. A prospective study of consecutively presenting diabetic patients admitted to the Department of Internal Medicine because of foot ulcer with a median follow-up of 4 years. Setting. A multidisciplinary foot-care team. Population. Five-hundred-and-fifty-eight consecutive diabetic patients with foot ulcers treated between 1 July 1983 and 3 1 December 1990 were followed to final outcome. Out of these patients, 468 healed either primarily (n = 345) or after minor or major amputations (n = 12 3) and 9 0 died before healing had occurred. Those 468 patients who healed were included in this prospective study from the time of healing. Main outcome measures. Patients were followed according to a standardized protocol with registration of foot lesions, amputation, morbidity and mortality. Clinical examination was performed twice yearly. Results. After 1 , 3 and 5 years of observation 34%. 61 % and 70% of the patients, respectively, had developed a new foot ulcer. The recurrence rate of foot lesions was slightly higher among patients who previously had had an amputation (P < 0.05, P < 0.0 1 and non-significant. respectively). Among patients with previous primary healing the cumulative amputation rates were 3%. 10% and 12% after 1 , 3 and 5 years of follow-up compared with 1 3 %. 3 5 % and 48 % among those who previously healed after amputation, irrespective of previous amputation level (P <: 0.001 at all time-points). All amputations except three were initiated by a foot ulcer deteriorating to deep infection or progressive gangrene. The long-term survival ratio was lower among patients healed after previous amputation (80%, 59%. 27%) compared with patients with previously primary healing (92%, 73%. 58%) after 1. 3 and 5 years of observation, respectively ( P < 0.001. P < 0.01 and P < 0.001 respectively). The mortality rate was twice as high among primarily healed and four times as high among patients with amputation compared to an age- and sex-matched Swedish population. Conclusion. These findings stress the need for life-long surveillance of the diabetic foot a t risk and the necessity of preventive foot care among diabetic patients with previous foot lesions, and particularly among those who had had a previous amputation.

Keywords: amputation, diabetes mellitus. foot ulcer, mortality.

Introduction Presence of foot lesions in diabetic patients is associated with development of deep infection and gangrene [ 1 , 21 which often require minor or major amputations [ 3-51. The short-term prognosis of these lesions is good if the patients are treated with a multidisciplinary team approach with early inter- vention [h . 71. The long-term prognosis is only

known from retrospective studies of diabetic patients with previous lower limb amputation [l. 3, 41. In those studies the 2-year mortality rate has been estimated to be 35-50% with a cumulative am- putation rate over 1-3 years of 40% [8-141.

The aim of the present study was to evaluate the long-term prognosis regarding recurrence of new foot ulcers, amputation and mortality in diabetic patients with previous foot ulcers treated and fol-

48 5

486 J. APELQVIST et al.

lowed in a prospective study by a multidisciplinary foot-care team.

Patients and methods

Patients

In a prospective study all diabetic patients who were referred to the Department of Internal Medicine between 1 July 1983 and 3 1 December 1990 because of foot ulcers were investigated (n = 558) and followed to final outcome both as in- and out- patients. Out of these 558 patients, 345 healed primarily, 12 3 healed after amputation (32 toes, 16 midtarsals, 63 below knee and 12 above knee) and 90 died before healing had occurred. From the time of healing (i.e. intact skin) those 468 patients were consecutively included in this prospective study and followed for a period of 6 months to 7 years (median 4 years). Thirteen were lost to follow-up (6 at 1 year, 2 at 2 years, 1 at 3 years. 2 at 4 years and 2 at 5 years of follow-up).

Clinical characteristics and definitions

At entry and during the course of the study the patients were treated by the same physicians. All data were collected using a preset standardized protocol. The most superficial lesion defined as an ulcer was a lesion through the full thickness of the dermis (Wagner grade 1) [ l ] . Wound classification, site of ulcer, presence of polyneuropathy, oedema. pain, distal perfusion pressure, albuminuria, retin- opathy. cardiovascular disease and living conditions were evaluated as has been previously described

The cause of death was established from clinical findings prior to death or by autopsy. The cause of death was divided into vascular and non-vascular.

Clinical signs of nephropathy were considered to be of diabetic origin when persistent albuminuria (three positive Albustixs) had been present for more than 6 months and if no other causes of kidney disease had been verified.

The retinal examination was performed through a

[ 15-1 81.

aneurysms, haemorrhages. hard exudates outside the macular region or with single ischaemic zones) and (c) severe retinopathy (preproliferative and proliferative retinopathy or previous photocoagula- tion). Seventeen patients were not examined due to their medical condition.

Smoking habits were graded into smokers (in- cluding those who had stopped smoking within 1 year prior to the study), previous smokers and non- smokers (defined as persons who never had had any daily nicotin consumption).

Medical and surgical treatment

All patients were treated and followed by the same foot-care team consisting of a diabetologist, an orthopedic surgeon, an orthotist, a podiatrist and a diabetes nurse. All patients were scheduled to an education programme in preventive foot care. The patients were followed with preset check-up visits, with clinical examination and laboratory evaluation, twice yearly. When the patients had a new foot lesion they were treated by the team both as in- and out-patients.

Metabolic control was improved when possible. Antibiotics, usually flucloxacillin, cefalosporins, metronidazole and oxikinolines, were used when clinical signs of infection (i.e. cellulitis. deep abscess or osteomyelitis) were present.

Femoral angiography and consultation with a vascular surgeon were performed in case of rest pain, progressive claudication, gangrene and in the pres- ence of a foot ulcer with a systolic toe pressure < 45 mmHg or an ankle pressure < 80 mmHg [h]. Surgical treatment (incision, drainage, bone resec- tion) and topical treatment and choice of foot wear have been previously described as the choice of analgesic agents (1 5. 16, 181. The lowest level con- sidered for amputation was at the interphalangeal level of the great toe. Strict amputation criteria were used for progressive gangrene (i.e. necrosis through all tissue levels), intolerable pain despite analgesia and septic or toxic conditions not responding to conservative treatment.

dilated pupil either by opthalmolscopy or fundus photography. The patients were examined within 1 Analytical techniques year of the time of enrolment. The most seriously affected eye was used for the evaluation. The patients Laboratory measurements were performed at ad- were classified into three categories depending on the mission and every 6th month during the observation degree of retinopathy. The groups were (a) no period. Further measurements were performed be- retinopathy, (b) simplex, (single or multiple micro- tween these prescheduled visits when new lesions

LONG-TERM OUTCOME IN DIABETIC FOOT ULCERS 487

Tablc 1 . I’iitient characteristic at inclusion

Primarily healed Amputated -

N 3 4 5 123 Age [years) 6 3 + 1 7 70 * 12t Sex (m/n 1 X2/163 67/56 Duration of diabetes (years) l h + l 2 1 7 k 1 3 Treatment 01)

Diet 37 1 1 Oral agents 9 7 41 Insulin 21 1 71

HbA,c (%) 8 . 7 + 2 . 1 8 . 3 + 1 . 9 Smoking habits (nl

Smokers/ex smokers 158 57 Non-smokers 1 H7 66

Systolic blood pressure (mmHg) 1 5 h + 2 4 1 6 3 k 2 5 Diastolic blood pressure (mmHg) 8 5 + 1 1 x 2 + 1 0 Systolic toe pressure (mmHg) 8 3 k 4 3 (39) 3 4 k 2 8 (17)f Systolic ankle pressure (mmHg) 1 3 2 k 4 7 (52) 9 0 k 4 4 ( 2 1 ) t Sensory neuropathy (VPT > H I ) (n) 231 (23) 104 (19)t Severe retinopathy (n) 1 x 1 (12) 58 ( 5 ) Diabetic nephropathy (Albustixa pos) (n) 72 39’

Mcansfsn ‘I’ c 0.05, ti’ < 0.001 vs. primarily healed. Numhcrs within bars indicate number of patients not investigated or with non-measurable bars.

had developed and then every 3rd month until healing had occurred. Glycosylated haemoglobin (HbA 1 c) levels were measured by ion-exchange chromatography using commercially available microcolumns (BIO RAD. Richmond, CA, USA). Normal value < 6.0%. Persistent proteinuria was measured with Albustix.’ (three measurements). Systolic toe and ankle blood pressure were measured with strain gauge and/or Doppler techniques as previously described [ h ] . The approximative mean of three measurements was used for the calculation. In 7 3 patients the ankle pressure was not measurable, usually because of incompressible arteries, and in 56 patients the toe pressure was not measurable because of site of ulcer or previous amputation. Sensory neuropathy was evaluated using a hand-held bio- thesiometer, (Bio Medical, Newbury. OH, USA) at the tip of the big toe as previously described [18, 191. A vibratory pressure threshold (VPT) > 30 arbitrary units (mean of three measurements) was considered as a sign of sensory neuropathy [ 191. Those patients unable to perceive any vibratory sensation were assigned a score of 5 1 [20]. VPT was not possible to measure in 42 patients due to their mental condition.

Statistical analysis

Values are given as means f SD. Differences between

or within groups were calculated using the Mann- Whitney U-test (two tailed) and the chi-squared test. In each group, survival was estimated by means of the Kaplan-Meier product limit estimator, of e.g. Altman [21], and confidence intervals were con- structed according to the logarithmic technique in Kalbfleisch & Prentice (221 : in these computations both age and sex were disregarded. Standardized mortality ratios were computed with the total Swedish population as reference: here age and sex were taken into account.

Results The patients included in the study were followed for a period of 6 months to 7 years (median 4 years). The clinical characteristics at entry for those patients with primary healing (n = 345) and patients who had healed after minor or major amputations (n = 123) are given in Table 1 . Patients who healed after an amputation were older ( P < 0.001), had lower distal perfusion pressure ( P < 0.001). more often had signs of sensory neuropathy ( P < 0.001) and neuropathy ( P < 0.05) compared with patients with primary healing. No differences were seen between the two groups regarding sex, duration of diabetes, treatment mode, metabolic control, smoking habits or blood pressure levels at inclusion. If signs of

488 J. APELQVIST et al.

100 -

90 -

80 -

70 -

60 -

50 - v

40 -

30 -

20 -

0 1 2 3 4 5 Years

Fig. 1. Cumulative recurrence rate of foot ulcers. Values are given as means (95% confidence limits). * P < 0.05: **P < 0 . 0 1 : m. amputated: 0. primarily healed.

muscular dysfunction (muscle wasting and atrophy), were included in the examination 96 % of the patients had severe signs of sensory or muscular dysfunction (n = 448).

Development of new foot ulcers

The cumulative recurrence rates of foot ulcers are shown in Fig. 1. Those patients with previous bilateral major leg (above ankle) amputation were excluded from the evaluation. The 1-. 3- and 5-year occurrence rates of foot ulcers among all patients were 34% 61 % and 70%. respectively. New ulcers were slightly more common among patients with previous amputation during the first 4 years of observation.

Y *** rr-

l o 0 / 0 I 2 3 4 5

Years

Fig. 2. Cumulative amputation rate. Values are given as means (95% confidence limit). ***P < 0.001 : .. amputated: 0, primarily healed.

Cumulative amputation rate

The cumulative amputation rate (both minor and major amputations) was estimated after exclusion of patients who died before entering the observed interval and those who previously had had a bilateral major amputation ( n = 10). All amputations except three were precipitated by a foot ulcer deteriorating to a deep infection or progressive gangrene. Those three major amputations were caused by proximal vascular disease ( n = 2) and as a complication after vascular surgery ( n = 1).

The cumulative amputation rates among all patients (Fig. 2) were 6%. 16% and 22 % after 1 , 2 and 3 years of observation. respectively. The cumu- lative amputation with previous primary healing

LONG-TERM OUTCOME I N DIABETIC FOOT ULCERS 489

I00

90

80

70

60

Y 7 50

40

30

20

10

0

I

!I* 0 I 2 3 4 5

Years Fig. 3. Cumulstive survival. Values are given as means (95% contidence limit). ***P < 0.001 : **I’ c 0.01 : .. amputated: 0. primarily healed.

Table 2. Standardized mortality ratios (SMR) in patient groups rclative to the total Swedish population during I-. 2- and 3- year periods of follow-up

1 year 2 years 3 years ~~

Primary healed 2 .04 2.OR 2.35 Amputated 4.99 4 .37 3.94

( P < 0.001, Fig. 2). In both groups there was no further increase in amputation rate after 4 years of observation. There was no difference in amputation rate among patients with previous minor amputation compared with those with a previous major am- putation (data not shown).

Cumulative mortality rate

The long-term survival rate was low among patients with a previous amputation (80% 59% 27%) compared with patients primarily healed (92 %- 73%. 58% after 1, 3 and 5 years of follow-up, respectively ( P -= 0.001, Fig. 3). Three out of four deaths were considered as vascular (data not shown). The dif- ference in mortality rate tended to increase by the number of observed years. The mortality ratios were twice as high among patients with primary healing and increased fourfold among patients with previous amputation compared to an age- and sex-matched Swedish population (Table 2).

Discussion This is the first major prospective study that shows a high risk for diabetic patients with previous foot ulcers to develop new lesions confirming previous observations [ 71, indicating that diabetic patients who once have developed foot ulcers always should be regarded as subjects at risk for developing new foot ulcers. This fact stresses the need for life-long preventive foot care and observation of the foot at risk. However, 30% of the patients did not develop a new ulceration during 5 years of observation indi- cating that recurrence of foot ulceration is possible to avoid.

The long-term prognosis for those who healed after an amputation was in poor agreement with previous studies in diabetic patients with lower limb amputation with a 3- and 5-year survival rate of only 65% and 41%, respectively [8-111. But the mortality was increased both among patients with primary healing and those who healed after am- putation compared to an age- and sex-matched Swedish population.

In the present study those who had undergone an amputation had a lower survival rate compared with those who had healed primarily. Further studies are needed to evaluate if this difference in survival rate can be explained by the differences in age, peripheral vascular disease and intercurrent diseases or by increased physiological stress imposed on the body by an amputation [23].

In the present study the 1- and 3-year amputation rates (15% and 35%) were in agreement with a previous observation by Jernberger [ 141, but in disagreement with other studies of diabetic patients who had had major amputation, with an average 1

490 J. APELQVIST et al.

to 3 year amputation rate of 42% [8. 10-1 31. The differences in cumulative amputation rate between different studies might be explained by differences in study design or population or by intensive multi- disciplinary treatment and approach to the diabetic patient and his foot at risk. The cumulative am- putation rate was much higher among patients with a previous amputation compared with patients with previous primary healing. The lower risk for am- putation among patients with primary healing might be explained by differences in age and multiple cardiovascular disease but amputation per se ir- respective of the level of surgical intervention might be a (independent) risk factor for further amputation. At the moment the number of patients surviving after amputation is too small and the observation period too short to draw any conclusions concerning that question from the present study.

The high risk for a future amputation among patients with a previous amputation has been suggested to be explained by an increased risk for developing foot ulcers [ l . 3. 41. However, in the present study recurrence of foot ulcers was only slightly more common among patients with previous amputation but almost all amputations in this study were precipitated by a foot ulcer. This indicates the need for an aggressive approach, particularly in patients with previous amputation, concerning pre- ventive foot care.

In conclusion, diabetic patients with previous foot ulcers run a high risk of developing new foot ulcerations, since 50% of the patients had developed a new ulcer within 2 years of observation. The cumulative amputation and mortality rates were especially high among those patients who had undergone amputation irrespective of amputation level. Almost all subsequent amputations were pre- cipitated by a foot ulcer and this finding stresses the need for life-long observation of the diabetic foot at risk and the necessity of preventive foot care es- pecially among diabetic patients with previous am- putation.

Acknowledgements This work was supported by the Swedish Research Council (grant no. 02872), the Hoechst Diabetes Foundation, the Swedish Diabetes Association and the Medical Faculty, University of Lund.

The cumulative analyses were performed by Pro-

fessor Jan Lanke of the Departments of Statistics and Mathematical Statistics, University of Lund.

References 1 k v i n ME. O'Neal LW. eds. The Diabetic Foot. 4th edn. St Louis:

Mosby. 19x8. 2 Frykberg RG. ed. The High Risk Foot in Diabetes Mellitus. New

York: Churchill Livingstone. 1991. 3 Bild DE. Selby JV, Sinnock P. Browner WS. Braveman P.

Showstack JA. Lower-extremity amputations in people with diabetes. Diabetes Cure 1989 : 12 : 24-30.

4 Fylling CP. Knighton DR. Amputation in the diabetic popu- lation : incidence. causes. cost, treatment and prevention. I Enterstom Ther 1989; 16: 247-55.

5 Most RS. Sinnock P. The epidemiology of lower extremity amputations in diabetic individuals. Diabetes Care 198 3 : 6:

h Apelqvist J. Castenfors J. Larsson J. Stenstrom A, Agardh C.-D. Prognostic value of systolic ancle and toe blood pressure levels in outcome of diabetic foot ulcer. Diabeks Care 19x9: 12: 373-x.

7 Edmonds ME, Blundell MP. Morris ME. Maelor Thomas E, Cotton LT. Watkins PI. Improved survival of the diabetic foot: the role of a specialized foot clinic. Q I Med 1 9x6 : 60 : 76 3-7 1 .

X Ecker ML. Jacobs BS. Lower extremity amputation in diabetic patients. Diabetes 1970: 19: 189-95.

9 Palumbo PI, Melton LJ. Peripheral vascular disease and diabetes. In: Diabetes in America: Diabeks data compiled 1 YX4. Washington. DC: US Govt Printing Otlice. 19x5. pp. 1-21 : NIH publication no. 85-14hX.

1 0 Silbert S. Amputation of the lower extremity in diabetes mellitus. Diabetes 1952: 1 : 297-9.

1 1 Cameron HC. Lennard-Jones JE. Robinson MP. Amputations in the diabetic outcome and survival. lancet 1964: 2 (Sept) 605-7.

12 Goldner M G The fate of the second leg in the diabetic amputee. Diabetes 1960; 9: 100-3.

1 3 Hoar CS. Torres 1. Evaluation of below-the-knee amputation in the treatment of diabetic gangrene. New Eng I Med 1962 : 266: 44(L3.

14 Jernberger A. Diabetesfoten. Swedish Medicinr (ISSN 02x4-

1 5 Apelqvist J , Castenfors J . Larsson 1. Stenstrom A. Agardh C.-D. Wound classification is more important than site of ulceration in the outcome of diabetic foot ulcers. Diabetir Med 1989 : 6: 526-30.

h Apelqvist J , Larsson J. Agardh C . 4 . The importance of peripheral pulses, peripheral oedema and local pain for the outcome of diabetic foot ulcers. Diabetic Medicine 1990: 7: 590-4.

7 Apelqvist J. Agardh C.-Il. The association between clinical risk factors and outcome of diabetic foot ulcers. Diab Rrs Cliri Pruct

1 X Apelqvist J. Larsson J , Agardh C-D. The influence of extenal percipitating factors and peripheral neuropathy on the de- velopment and outcome of diabetic foot ulcers. j Iliabefir Compl 1990: 4 : 21-5.

19 Boulton AJM. Kubrusly DB. Bowker JH. Gardia MT. Uuintero L. Becker DM. Skyler JS. Sosenko IM. Impaired vibratory perception and diabetic foot ulceration. Diabetic Med 19x6: 3 : 335-7.

X7-91.

5342) 1989: 14: 176-83.

1992; 18: 43-53.

LONG-TERM O U T C O M E I N D I A B E T I C F O O T ULCERS 491

2 0 Houlton AJM. Hardisty CA. Betts RP, Franks CI. Worth RC. Ward JD. Duckworth T. Dynamic foot pressure and other

neuropathy. Uiaktes Carp 1983: 6: 26-3 3. 2 1 Altman DG. Prartiral Statistics /or Mediral Research. London:

Chapman and Hall. 199 1. 22 Kalbfleisch ID. Prentice RIA. This Statistiral Analgsis of Failure

Time Data. New York: Wiley and Sons, 1980.

23 Sims D. Cavanagh PR, Ulbrecht JS. Risk factors in the diabetic foot-recognition and management. Physical Therapy 1988 :

studies as diagnostic and management aids in diabetic 68: 1887-1902.

Received 1 September 1992. accepted 20 November 1992.

Correspondence: Jan Apelqvist MD. Department of Internal Medi- cine, IJniversity Hospital, S-221 85 Lund. Sweden.

I M H 233