Epidemiology of Nontraumatic Lower-Extremity Amputation in Area 7, Madrid, Between 1989 and 1999: A population-based study

Diabetes remains the main cause of lower-extremity amputations (LEAs) (1,2,3). Recently (3), the number of LEAs carried out worldwide has been estimated to be >162 × 10⁶ each year, with a high economic and social cost. More than 50% of these LEAs are performed in subjects with diabetes. Differences in the incidence of LEAs could reflect variations in risk factors, diabetes prevalence, health care systems, and lifestyle, as well as differences in population age-structure, the sources from which the case was identified, the levels of ascertainment, and the definitions of LEAs and the distinctions between major and minor LEAs. Recently, a standard approach to data collection has been described (4,5), and the first article (3) about the epidemiology of LEAs in several countries has been published. In this study, area 7 from Madrid presents the lowest LEA rates. The purpose of this article is to compare the incidence of LEAs in diabetic and nondiabetic subjects—in relation to the provision of improved foot-care after the St. Vincent Declaration—by use of a standardized approach, in the area 7 of Madrid between 1989 and 1993, between 1994 and 1996, and between 1997 and 1999.

The National Health Service has divided Madrid into 13 health care areas. The public health care system is serving >99% of the total population from the area 7 and is provided by a single hospital (Hospital Clínico San Carlos) with a single specialized service for vascular patients. According to the census in 1991, the total population for the catchment area was 569,307 (261,529 men and 307,778 women). The population with diabetes was calculated according to the Lejona study (6). In this study, the prevalence of diabetes was estimated by a randomized sample stratified by 5-year age-groups and sex. The data of this study were used to calculate the number of people with diabetes according to age and sex distribution, bearing in mind that the population composition in both Lejona (Vizcaya) and Madrid are comparable. The total diabetic population estimated (diagnosed and unknown) was 37,932 (15,505 men and 22,427 women). From 1991 to 1996, the date of the last census, the catchment area was reduced by ∼4,000 inhabitants, which was taken into account in order to calculate incidence rates. According to the Global Lower Extremity Amputation Study (3), amputation was defined as the complete loss of any part of the lower limb. The term “minor amputation” refers to an amputation distal to the tarsometatarsal joint, whereas “major amputation” refers to one through or proximal to the tarsometatarsal joint. All patients who underwent an LEA between 1 January 1989 and 31 December 1999 were identified through operating theater records. Vascular surgery department and endocrinology service discharge records were used as secondary sources. Only patients who had resided in area 7 for at least the last 6 months were included in the study. The social security system in our country covers between 70 and 100% of the cost of pharmacological treatments when prescribed by a social security doctor. Patients suffering from LEAs usually have some pharmacological treatment. Thus, prescribing physicians (family doctors and physicians of the endocrinology service in an outpatient setting) were used as an additional source to identify people who were amputated outside area 7. This source was operative since 1994. Date of birth, sex, address, date and level of the actual amputation, and potential causes of the actual amputation were registered. Estimations of the level of ascertainment were derived by using capture-recapture methods (5). Because three independent sources were available from 1994, the level of ascertainment was calculated during the 1994–1999 period. To have a greater number of cases, men and women were analyzed together.

Incidence of LEA data are presented over three periods. From 1989 to 1993 no intervention was done. Since 1994, a series of improvements in therapeutic measures in diabetes management are available in our area, including a prophylactic foot care teaching and treatment program (7). As the efficacy of these measures have improved across time, we have studied the incidence of LEAs during two 3-year periods (1994–1996 and 1997–1999). From 1989 to 1993, 139 diabetic subjects and 118 nondiabetic subjects underwent 156 and 139 nontraumatic LEAs, respectively, while from 1994 to 1999, 91 diabetic and 43 nondiabetic subjects suffered 111 and 49 nontraumatic LEAs, respectively. Age-adjusted incidence per 10⁵ people in risk per year for first LEAs (95% CI) decreased from 1989–1993 to 1997–1999 in diabetic people as follows: major LEAs decreased from 67.1 (60.9–73.3) to 12.3 (14.1–10.5) and from 13.3 (11.6–15.0) to 5.6 (4.9–6.3) for men and women, respectively; minor LEAs decreased from 52.1 (45.0–59.2) to 22.5 (19.7–25.3) and from 10.9 (10.3–11.5) to 7.9 (6.8–9.0) for men and women, respectively. Data for nondiabetic people are as follows: major LEAs decreased from 2.6 (2.2–3.0) to 1.1 (0.4–1.8) and from 1.3 (1.1–1.5) to 0.5 (0.4–0.6) in men and women, respectively; minor LEAs decreased from 2.7 (2.1–3.3) to 0.7 (0.6–0.8) and from 1.5 (1.3–1-7) to 0 for men and women, respectively. Crude and age-adjusted incidence per 10⁵ people in risk per year for first and all (major and minor) LEAs for the three periods are displayed in Table 1.

The incidence of LEAs rose steeply with age >40 years and it rose dramatically with age >80 years; it was higher in men than in women. The incidence of major LEAs was greater than that of minor LEAs in nondiabetic people, whereas diabetic subjects more frequently suffered minor LEAs.

For major LEAs, estimations of ascertainment were 100 and 98.7%, and for minor LEAs, they were 83.8 and 76.9%. Given that more than one condition was present per patient, LEAs in diabetic subjects were associated with peripheral vascular disease (PVD) (major/minor: 100/62%), neuropathy (major/minor: 78/92%), and infection (major/minor: 24/84%), whereas LEAs in nondiabetic subjects were associated with PVD (major/minor: 100/98%), neuropathy (major/minor: 22/33%), and infection (major/minor: 6/18%). According to the St. Vincent Declaration, several studies in recent years have found a decrease in the incidence of LEAs in people with diabetes.

However, only a few studies (8,9,10) have analyzed a 10-year period, as this one has. Furthermore, to our knowledge, no other study has used the capture-recapture method to estimate the level of ascertainment from more than two separate data sources and to measure the incidence of first and all LEAs in a separate way in the whole population (i.e., those with and without diabetes). According to our data, LEA incidence in area 7 of Madrid remains the lowest reported incidence in a Caucasian population of both diabetic and nondiabetic people. Primary prevention is defined as a reduction in the incidence of a disease, e.g., first LEA. Since 1994, following the provision of improved foot care after the diabetes program was available (7), a reduction in first LEA of 57 and 81% (minor and major, respectively) in diabetic men and of 28 and 57% were detected, with a 5- to 6-year increase in age at the first LEA.

These data indicate that efforts to delay and reduce the incidence of first LEA in people with diabetes succeeded. In a similar way, a slightly higher reduction in first LEAs was found in the nondiabetic population also at the end of the study. Several factors may contribute to these data. The decline in LEA rates can be partially explained by favorable trends in some risk factors and a better control in others, including dietary factors as well as a tighter control of hypertension, hypercholesterolemia, and reduced tobacco and alcohol consumption (11). Better control of these risk factors should positively affect the entire population. In addition, the increasing availability of resources for the usual clinical management of these patients, including new drugs and surgical procedures (12) covered by the social security system, should also be operative.

This study also evaluated the conditions associated with LEAs. The proportion of major LEAs associated with PVD was high and similar in both diabetic and nondiabetic population. The proportion of minor LEAs associated with neuropathy and infection in diabetic subjects was clearly greater than that in the nondiabetic population. This could explain the excess of minor LEAs in diabetic people, suggesting that we should try to cope with these problems.

In conclusion, our data show that the incidence of LEAs in area 7 of Madrid remains as the lowest incidence reported in European countries in both diabetic and nondiabetic people. A substantial decrease in LEAs and a later presentation in relation to a series of improvements in diabetic treatment were detected. Despite these figures, the incidence of LEAs remains higher in diabetic subjects than in the nondiabetic population, suggesting that diabetic foot care remains suboptimal in Madrid. Taking into account that diabetic subjects suffer minor LEAs more frequently than nondiabetic people and that it is associated with neuropathy and infection, a more substantial reduction in LEAs in diabetic people should be achieved with an earlier neuropathy diagnosis and an adequate antibiotic management.