ARTICLE. Björn Rathsman & Kerstin Jensen-Urstad & Thomas Nyström

Transcription

1 DOI /s ARTICLE Intensified insulin treatment is associated with improvement in skin microcirculation and ischaemic foot ulcer in patients with type 1 diabetes mellitus: a long-term follow-up study Björn Rathsman & Kerstin Jensen-Urstad & Thomas Nyström Received: 9 January 2014 /Accepted: 3 April 2014 # Springer-Verlag Berlin Heidelberg 2014 Abstract Aims/hypothesis We investigated skin microcirculation and its association with HbA 1c and the incidence of ischaemic foot ulcer in patients with type 1 diabetes formerly randomised ( ) to intensified conventional treatment (ICT) or standard treatment (ST) with insulin for a mean of 7.5 years. Methods We re-determined the skin microcirculation of 72 patients (ICT 35 vs ST 37) from the original Stockholm Diabetes Intervention Study with iontophoresis topically applied with the following vasoactive stimuli: acetylcholine (ACh) (endothelial-dependent vasodilatation), sodium nitroprusside (SNP) (endothelial-independent vasodilatation), and capsaicin (C-nociceptive-dependent vasodilatation). HbA 1c levels (mean of 14 values/patient) were prospectively collected between 1990 and 1995 and tested for association with skin microcirculation. The patients were followed until first hospitalisation for an ischaemic foot ulcer or until Results During the median 28 years of follow-up, three patients developed ischaemic foot ulcers in the ICT group compared with ten in the ST group (logrank, p=0.035). At the time of iontophoresis, HbA 1c was lower in the ICT group (median 57 mmol/mol [minimum maximum mmol/mol]) compared with the ST group (68 mmol/mol [41 96 mmol/mol], p<0.01) (DCCT: ICT B. Rathsman Department of Clinical Science and Education, Sachs Children s Hospital, Karolinska Institutet, Stockholm, Sweden K. Jensen-Urstad Section of Clinical Physiology, Karolinska Institutet, Stockholm, Sweden T. Nyström (*) Division of Internal Medicine, Department of Clinical Science and Education, Karolinska Institutet, Södersjukhuset, Stockholm, Sweden thomas.nystrom@ki.se 7.4% [ %] vs ST 8.4% [ %]). Stimulated blood flow was higher in the ICT vs ST group with significantly increased perfusion units (PU) for: ACh (8.1 PU [ PU] vs 5.3 PU [ PU], p<0.01); SNP (8.1 PU [ PU] vs 5.6 PU [ PU], p=0.03); and capsaicin (5.0 PU [ PU] vs 3.4 PU [ PU], p<0.01). HbA 1c was associated with vasodilatation induced by ACh (b= 0.02, p<0.01) and capsaicin (b= 0.02, p=0.03). HbA 1c was independently associated with ACh (b= 1.48, p<0.01) and capsaicin-induced vasodilatation (b= 1.45, p<0.01). Conclusions/interpretation Improved glycaemic control in patients with type 1 diabetes is associated with an improvement in skin microcirculation and with a lower incidence of ischaemic foot ulcers. Trial registration: ClinicalTrials.gov NCT Keywords HbA 1c. Intensified insulin treatment. Ischaemic foot ulcer. Skin microcirculation. Type 1 diabetes Abbreviations ACh Acetylcholine ICT Intensified conventional treatment SNP Sodium nitroprusside SDIS Stockholm Diabetes Intervention Study ST Standard treatment Introduction Foot ulcers are common in patients with diabetes, sometimes leading to serious complications including leg amputation [1]. A risk factor for the development of foot ulcers may be poor glycaemic control as reflected by high levels of HbA 1c [2]. In long-standing hyperglycaemia, the endothelium of blood vessels is adversely affected [3]; endothelial

2 dysfunction, common in diabetes patients, is one of the earliest steps in the development of atherosclerosis and occurs in both small and large vessels. A linkage between such micro- and macrovascular complications has been observed, with microvascular endothelial dysfunction often preceding large-artery stiffening and cardiovascular complications [4]. Microangiopathy is not the only or main cause of foot ulcers, and several other risk factors have to be considered [5]. For example, neuropathy is a main contributor to foot ulceration [5]. Diabetic neuropathy and endothelial dysfunction are closely associated, demonstrated by the fact that a normal neurogenic vascular response, mediated by stimulation of C-nociceptive nerve fibres, contributes to one-third of the maximal vasodilatory capacity under conditions of stress [6]. Improved glycaemic control by intensified insulin treatment in type 1 diabetes halts the development of microvascular complications [7, 8] and has long-term beneficial effects on macrovascular complications [9], probably also as the result of improved glycaemic control [10]. Because of the beneficial effects of improved glycaemic controls by intensive insulin treatment on the micro- and macrovascular systems of type 1 diabetes patients, we hypothesised that patients who had been randomised to intensive insulin treatment would also have improved skin microcirculation and a decreased incidence of ischaemic foot ulcers. To test this, we compared the skin microcirculation and the first time of hospitalisation for ischaemic foot ulcer in type 1 diabetes patients from the Stockholm Diabetes Interventional Study (SDIS) cohort [8]. These patients were formerly randomised ( ) to intensified conventional treatment (ICT) or standard treatment (ST) with insulin for an average of 7.5 years (until 1990). Methods Participants The original SDIS aimed to determine whether intensified insulin treatment in type 1 diabetes improved diabetic microvascular complications [8]. A total of 102 patients with type 1 diabetes mellitus were randomised to ICT (n=48) or ST (n=54) with insulin between September 1982 and March The treatment mode and the inclusion criteria for the original SDIS study have been described elsewhere [11]. Briefly, the ST patients were advised to monitor their blood glucose and the treatment goal was to reduce blood glucose values without giving rise to hypoglycaemia. Participants visited the clinic every fourth month. The patients in the ICT group participated in a structured educational programme and were recommended multiple insulin injections and frequent home blood-glucose monitoring. They visited the clinic bimonthly. The original SDIS study lasted for an average of 7.5 years and ceased in 1990, by which time 96 patients had been fully evaluated [8]. Thereafter, all patients were assigned to their regular clinical visits, mainly at one hospital (Södersjukhuset, Stockholm, Sweden). Eventually, in accordance with guidelines established for the treatment of type 1 diabetes [7, 8], all patients were recommended by their regular physician to have multiple insulin injections (i.e. pre-meal and basal insulin therapy). For the present SDIS iontophoresis study, 96 patients from the original SDIS study [8] were asked to have their skin microcirculation investigated by iontophoresis; 72 patients agreed (ICT 35 vs ST 37). Exclusion criteria were: any history or ongoing ischaemic foot ulcer or foot and/or leg amputation; osteoarthropathy; abuse of alcohol or drugs; or mental illness. However, none of the patients asked to participate was excluded based on these criteria. The iontophoresis procedure took place in The controls for the iontophoresis method used in the study were 19 healthy individuals. All 72 participants (except for the healthy controls) were then prospectively followed until the first time of hospitalisation for ischaemic foot ulcer or until the censor date of 31 December A flow chart for the study is given in Fig. 1. The present study was approved by the regional ethics committee and complied with the Declaration of Helsinki. All the patients gave their informed consent before participating. This trial is registered (ClinicalTrial.gov registration no. NCT ). Clinical data Clinical data such as BMI, tobacco use, blood pressure (systolic and diastolic), insulin dosing, anti-hypertensive medication, lipid-lowering agents and anti-platelet therapy were obtained at the time of iontophoresis and retrospectively (at 2011) from medical journals and the Swedish National Diabetes Register [12]. HbA 1c and total cholesterol HbA 1c was studied as previously described [8]. Briefly, in the original SDIS study, HbA 1c was measured at entry, after 6 months and then approximately at every fourth month in both groups. HbA 1c (mean of 22 values/ patient) was calculated for the whole 7.5 year period in the original SDIS study [8]. In the present study HbA 1c levels were prospectively collected between 1990 and HbA 1c (mean of 14 values/ patient) was calculated for that period and tested for association with skin microcirculation. Thereafter, HbA 1c levels were retrospectively collected between 1996 and 2010 (mean of 20 values/patient) and evaluated in 3 year intervals ( , , , and ) and finally the last year of the follow-up in 2011 (one to three values/patient). Blood lipids were not evaluated in the original SDIS study [8]. In the present study, we retrospectively collected data for total cholesterol level. Total cholesterol levels (mean of ten values/patient) were evaluated in the same 3 year

3 Enrolment ICT (n=48) Enrolment Assessed for eligibility (n=121) End of original SDIS (1990) Lost to follow-up (n=6) due to moving from Stockholm, Sweden Formerly allocated to ICT (n=35) Analysed (n=35) Original SDIS Randomised (n=102) ( ) Allocation Follow-up SDIS iontophoresis (1995) Assessed for eligibility Randomised (n=96) Follow-up Analysis Excluded (n=19) Declined to participate (n=19) ST (n=54) Excluded (n=24) Declined to participate (n=17) Other reason (n=7) Formerly allocated to ST (n=37) Analysed (n=37) Fig. 1 Flow chart for the SDIS cohort study. In the original SDIS study patients were recruited between 1982 and 1984 and randomised to ICT or ST with insulin for a mean of 7.5 years. In 1990, the original SDIS study ceased and 96 patients were fully evaluated [2]. Thereafter, patients were treated routinely by their clinical diabetes team, most by multiple insulin injections (i.e. pre-meal and basal insulin therapy). In 1995 (5.5 years after the SDIS ceased) patients were asked to participate in the present SDIS iontophoresis study and have their skin microcirculation investigated. Thereafter, patients were prospectively followed until hospitalisation for ischaemic foot ulceroruntilthecensordatein December HbA 1c levels were collected prospectively between 1990 and 1995 (mean of 14 values/patient) when iontophoresis was conducted. Between 1996 and 2011, HbA 1c levels (mean of 20 values/ patient) and total cholesterol levels (mean of ten values/patient) were collected retrospectively intervals as for HbA 1c. Total cholesterol levels (one to two values/patient) were also evaluated at the end of the follow-up in Microvascular complications The results from the original SDIS study regarding microvascular complications have been published elsewhere [8] and were not the outcome of the present study. As microvascular complications strongly associate with macrovascular complications, the incidence of severe microvascular complications at the time of iontophoresis was used in the linear regression model against the measurements of skin microcirculation. The assessments of severe microvascular complications have also been described elsewhere [8]. Briefly, severe retinopathy was defined as any proliferative retinopathy or macula oedema requiring photocoagulation based on fundus examinations [8]. Severe nephropathy was defined as albuminuria greater than 200 μg min 1. Severe peripheral neuropathy was defined as reduced nerve conduction velocity in at least one nerve [8]. Iontophoresis and ankle brachial index Studies were conducted between breakfast and lunch in a quiet room kept at C. Patients took their ordinary morning dose of insulin and had a standard breakfast and were asked to refrain from smoking or drinking coffee. During iontophoresis, the patients were in a supine position with their feet at heart level. First, blood pressure was measured sphygmomanometrically in the right arm. The ankle pressure was obtained with a Doppler signal followed by release sphygmomanometrically on the dorsal right foot. A probe holder for combined single-point iontophoresis and blood flow measurement was fixed on the dorsum of the right foot. Skin blood flow at a depth of approximately 1 mm was registered by a laser Doppler flowmetry technique (Periflux, Perimed, Stockholm, Sweden). Basal flow was calculated as an average of flow for 10 s during the last minute before the start of iontophoresis. Blood flow after administration of drugs was defined as the average blood flow for the 10 s with the maximal flow following the stimulation. Results were given as the change in blood flow for each of the vasoactive substances used, calculated as a ratio between the maximum blood flow after stimulation and basal flow, and presented as arbitrary perfusion units (PU). The chamber holder was filled with 2% (wt/vol.) ACh (Dispersa Hettlingen, Zürich, Switzerland) solution in deionised sterile water. Iontophoresis was performed with a 0.20 ma current for 15 s and repeated with a longer stimulation time (30 s) and maximal stimulation of s. At a different site, on the same foot, and not simultaneously, 2% (wt/vol.) SNP (Nipride, Hoffman-La Roche, Basel, Switzerland) and 2% (wt/vol.) capsaicin (Sigma-Aldrich, St Louis, MO, USA) in solutions in deionised sterile water were iontophoresed at 0.5 ma to obtain a dose response curve with the same time points as followed previously. For the ACh we used an anodal current, while for the SNP and the capsaicin we used a cathodal current. The investigator and the interpreter of the results were blinded to the former allocation of the individual patient.

4 Ischaemic foot ulcer The follow-up study outcome was a first hospitalisation for ischaemic foot ulcer, defined by discharge codes in the Swedish National Inpatient Register: ICD , , , , and ( default.htm); and ICD-10 E10.5, E 10.6, E10.7, I70.2 and L98.4, ( This register has covered the whole country since 1987 [13]. The diagnostic validity for ischaemic diabetic foot ulcers in the register has been shown to be 98% if it is the principal cause of hospitalisation [14]. Patients were followed from the time of inclusion in the original SDIS until the primary diagnosis of ischaemic foot ulcer or the end of follow-up at the censor date of 31 December All cases (n=13) were then confirmed by two clinicians using medical journals with reference to the diagnostic criteria for ischaemic foot ulcer, which was defined by ongoing foot ulcer (primary) together with one of the following (secondary): ankle brachial index <0.9 or toe pressure <50 mmhg, with a slight modification, according to the International Working Group of the Diabetic Foot [15]. Statistical analyses Results are presented as median (minimum maximum). Comparisons between treatment groups for continuous variables were made with the Mann Whitney U test. Contingency tables were analysed using the χ 2 test. Associations were determined with linear univariable and multiple regression analysis. The measurements of skin microcirculation for each of the vasoactive substances used (ACh, SNP and capsaicin) were log transformed in order to achieve normality and then studied in the regression analyses for predefined factors known to be associated with endothelial function. Information on these factors was collected at the time of the iontophoresis (1995): age; duration of diabetes; HbA 1c ; smoking; systolic blood pressure; and severe microvascular complications (i.e. retinopathy, nephropathy and neuropathy). Time to first hospitalisation for ischaemic foot ulcer was presented using the Kaplan Meier method, and differences between ICT and ST groups tested with the logrank test. The result was considered significant if p<0.05. All statistical analyses were performed using PASW/SPPSS Statistics 18 software package (SPSS, Chicago, IL, USA). Table 1 Clinical characteristics of the SDIS cohort at the time of iontophoresis (SDIS iontophoresis 1995) Characteristic SDIS iontophoresis (1995) End of follow-up (2011) ICT ST ICT ST n (women/men) 35 (21/14) 37 (22/15) 32 (19/13) 27 (16/11) Age (years) 42 (28 63) 42 (31 63) 56 (45 80) 58 (49 69) Duration of diabetes (years) 28 (19 45) 27 (19 39) 44 (36 61) 44 (37 56) BMI (kg/m 2 ) 24(19 30) 24 (19 30) 25 (18 36) 25 (17 34) Current cigarette smokers (n) Blood pressure (mmhg) Systolic 135 ( ) 135 ( ) 130 ( ) 135 ( ) Diastolic 77 (60 100) 75 (60 95) 70 (58 85) 70 (60 90) HbA 1c (mmol/mol) 57 (40 79) 68 ** (41 96) 63 (51 90) 67 (41 98) HbA 1c (%) 7.4 ( ) 8.4 ** ( ) 7.9 ( ) 8.3 ( ) Total cholesterol (mmol/l) 4.5 ( ) 5.1 ( ) Nephropathy (n) 2 8 * Retinopathy (n) ** Peripheral neuropathy (n) 5 10 Insulin dose (U kg body weight 1 day 1 ) 0.7 ( ) 0.7 ( ) 0.6 ( ) 0.6 ( ) Anti-hypertensive medication (n) * Lipid-lowering agent (statin) (n) Anti-platelet therapy (n) The iontophoresis study took place 5.5 years after the original SDIS stopped. Data on patient characteristics were again collected at the end of follow-up in 2011 (median follow-up of 28 years) HbA 1c and total cholesterol are median values (minimum maximum), from the end of the original SDIS study Nephropathy is defined as albuminuria >200 μg/min; retinopathy is defined as any proliferative retinopathy or macula oedema requiring photocoagulation based on fundus examinations; peripheral neuropathy was defined as reduced nerve conduction velocity in at least one nerve * p<0.05 and ** p<0.01, ST vs ICT, at a given time point

5 a HbA 1c (mmol/mol) b Total cholesterol (mmol/l) Results SDIS baseline SDIS baseline End of SDIS End of SDIS * ** *** SDIS iontophoresis SDIS iontophoresis End of follow-up End of follow-up Fig. 2 (a) Mean HbA 1c levels from the SDIS cohort study with a median follow-up of 28 years. *p<0.05, **p<0.01 and ***p<0.001, between ICTand ST at given time points. (b) Total cholesterol was not measured at the beginning of SDIS. Mean total cholesterol levels were retrospectively collected from 1996 until 2011 ( end of follow-up ). White circles, ICT; black squares, ST Clinical data Characteristics for the study time period are presented in Table 1. Group characteristics showed no statistically significant differences, except for a significantly higher use of anti-hypertensive treatment in the ST group at 2011 (Table 1). In accordance with the original SDIS results [8] the incidence of nephropathy and retinopathy in the ICT group were significantly lower at the time of iontophoresis (Table 1). HbA 1c and total cholesterol HbA 1c levels and cholesterol levels are shown in Table 1 and Fig. 2a, b. Levels of HbA 1c did not differ between groups according to the time of randomisation (SDIS baseline: ICT 76 mmol/mol [ mmol/mol] vs ST 79 mmol/mol [ mmol/mol]; DCCT, ICT 9.1% [ %] vs ST 9.4% [ %], p=0.66 vs patients included in current study SDIS baseline: ICT 76 mmol/mol [ mmol/mol] vs ST 79 mmol/mol [49 99 mmol/mol]; DCCT, ICT 9.1% [ %] vs ST 9.4% [ %]). The retrospective data for HbA 1c and total cholesterol from 1996 to 2010 are shown in Fig. 2a, b.hba 1c levels were very similar between the groups 20 years after inclusion in the original SDIS study (Fig. 2a). Total cholesterol levels did not differ between groups from 1996 to 2010 (Fig. 2b), or at the end of the study (Table 1). Ischaemic foot ulcer incidence rate During the median 28 years of follow-up, 13 patients developed ischaemic foot ulcer. Three patients out of 35 in the ICT group with a median follow-up of 29 years developed an ischaemic foot ulcer during follow-up, compared with ten out of 37 in the ST group with a median follow-up of 28 years (logrank test p=0.035) (Fig. 3). Iontophoresis and ankle brachial index Ankle brachial index did not differ between groups (ICT 1.1 PU [ PU] vs ST 1.1 PU [ PU], p=0.54). The skin microcirculation data are presented in Table 2. Basal skin microcirculation blood flow did not differ between groups (ICT 4.5 PU [ PU] vs ST 5.9 PU [ PU], p=0.33). As expected, healthy participants had a higher skin microcirculation for all vasoactive substances used in the method compared with the diabetes patients regardless of former randomisation (Table 2). The skin microcirculation blood flow was higher in the ICT group compared with the ST group for all three vasoactive substances applied (Table 2). The differences between groups were already evident during the shortest stimulation time (15 s) when ACh or SNP was applied, and with the second shortest stimulation time (30 s) during iontophoresis with capsaicin-induced vasodilatation (Table 2). Association between HbA 1c and skin microcirculation For the predefined factors used in the linear regression analysis against skin microcirculation (for the maximal stimulation) a negative association was seen between mean levels of HbA 1c ( ) and vasodilatation induced by ACh (b= 0.02, p<0.01) and capsaicin (b= 0.02, p=0.03), with a borderline association with SNP-induced vasodilatation (b= 0.01, p=0.052). There was also a negative association between Disease-free survival Observation time (years) Fig. 3 Kaplan Meier disease-free survival curves for ICT and ST group patients. p= Dashed line, ICT; solid line, ST 25 30

6 Table 2 Skin microcirculation data (iontophoresis) for each of the vasoactive drugs applied Drug severe retinopathy and ACh-induced vasodilatation (b= 0.23, p=0.047). By using the other two time points, i.e. 15 s and 30 s, in the linear regression analysis the results did not change (data not shown). In the linear multiple regression analyses adjusting for age, duration of diabetes, smoking, systolic blood pressure and severe microvascular complications, with the vasoactive substances (ACh, SNP and capsaicin) as dependent variables and HbA 1c as the explanatory factor, HbA 1c was independently associated with vasodilatation induced by ACh (b= 1.48, p<0.01) and capsaicin (b= 1.45, p<0.01). Discussion SDIS iontophoresis ICT (n=35) ST (n=37) p value Healthy controls (n=19) ACh (PU) 15 s 2.9 ( ) 2.1 ( ) s 5.1 ( ) 3.5 ( ) ( ) Max. 8.1 ( ) 5.3 ( ) SNP (PU) 15 s 4.1 ( ) 2.6 ( ) s 6.5 ( ) 4.6 ( ) ( ) Max. 8.1 ( ) 5.6 ( ) Capsaicin (PU) 15 s 1.4 ( ) 1.3 ( ) s 1.8 ( ) 1.5 ( ) ( ) Max. 5.0 ( ) 3.4 ( ) The three vasoactive drugs were: ACh (endothelial-dependent); SNP (endothelial-independent); and capsaicin (C-nociceptive-dependent) In addition, 19 healthy individuals were investigated as controls Iontophoresis values are median (minimum maximum) The iontophoresis skin blood flow was calculated as a ratio between 15 s, 30 s and maximal flow, respectively, against the basal flow Max., maximal stimulation between 60 and 120 s We here demonstrate that patients in the SDIS formerly randomised to intensive conventional insulin treatment for a mean of 7.5 years have a lower incidence of ischaemic foot ulcer during long-term follow-up (median 28 years) compared with patients randomised to standard insulin treatment. The intensified-insulin-treated group also had a better skin microcirculation response and lower HbA 1c levels compared with the standard-insulin-treated group (at 5.5 years after the original SDIS ceased). In a multiple regression analysis, HbA 1c was negatively associated with skin microvascular function. The SDIS and DCCT were the first two randomised studies demonstrating that intensified insulin treatment in type 1 diabetic patients retards microvascular complications [7, 8]. In the much larger DCCT study there was a borderline significant decrease in the number of macrovascular events in the intensively treated group [7], which became statistically significant some years later [9]. Previous reports from the SDIS cohort have demonstrated that early atherosclerosis can be retarded by intensified insulin treatment [16]. Although the development of atherosclerosis is multifactorial, there is evidence indicating chronic hyperglycaemia as a culprit in macrovascular complications [9, 16]. If chronic hyperglycaemia is treated in the early stages of type 1 diabetes, the progression of atherosclerosis can be halted [9], which also has been observed in type 2 diabetes [17]. In those studies, the beneficial action on the macrovascular outcomes became obvious later during follow-up, at the same time as the levels of HbA 1c became similar between groups [9, 17]. It has therefore been suggested that a glucose memory exists and that this glucose memory protects against later macrovascular complicationsobservedinpatientswithdiabetes[9, 17]. In the present study, the glucose control (mean HbA 1c ) was still slightly different between groups 5.5 years after the cessation of the original SDIS study and did not entirely converge until 20 years of follow-up. This may have further protected these patients against hyperglycaemiainduced vascular dysfunction. We also confirmed that skin microcirculation in the foot is impaired in patients with type 1 diabetes compared with healthy controls, in accordance with recent studies [18]. However, the intensified-insulin-treated patients had better skin microcirculation for all of the three substances used in the iontophoresis procedure compared with the standard-insulintreated patients. Except for the long-standing hyperglycaemia no other distinctions in risk factors affecting microcirculation were observed between groups. This may suggest that chronic hyperglycaemia causes a deterioration in skin microcirculation. In the current study, endothelial-dependent skin vasodilatation was associated not only with HbA 1c but also retinopathy. There is a close association between endothelial dilatory response in the skin microcirculation and the large arteries; changes in skin microcirculation and macrovascular function of the large vessel are important for the development of ischaemic foot ulcers [18, 19]. The microvascular complications neuropathy [19] and nephropathy [20] are both independent risk factors for macrovascular disease. However, after introducing the predefined factors in the multiple regression model, only HbA 1c was associated with endothelialdependent and C-nociceptive-dependent induction of skin microvascular blood flow, suggesting that these risk factors (including severe microvascular complications) did not greatly affect skin microcirculation. The neurovascular response mediated by C-nociceptive nerve fibres is of some importance

7 to skin microcirculation [6]. Stimulation of these fibres releases vasodilating substances, such as bradykinin, in a manner relating to nerve axon reflex-related vasodilatation, which accounts for up to one-third of the total ACh-induced skin microcirculation. This is in contrast with SNP-induced skin microvascular vasodilatation, which works through completely different pathways from ACh induction of skin microvascular vasodilatation [21]. This might explain the observed associations between mean HbA 1c levels and both endothelial- and capsaicin-dependent skin vasodilatation, with no such association between HbA 1c and SNP-induced vasodilatation. The strength of our study includes the long-term follow-up. Limitations of the study include the small size of the SDIS cohort study and that the present study was not initially powered to investigate macrovascular complications [16]. We have no baseline data for the skin microcirculation, which might have differed between groups at that time, though the randomisation was otherwise successful [11]. A potential moderating factor of skin microcirculation assessed by iontophoresis is electrical skin resistance, which probably has a complex relationship with blood flow [22, 23]. Diabetes may be associated with high electrical skin resistance [22], and age is also associated [24], which might explain the differences between diabetic patients and healthy controls, albeit not between the diabetic groups. Strong risk factors for macrovascular complications, such as different lipoproteins, were not measured in the initial part of this cohort, nor do we have any total cholesterol levels or data for anti-hypertensive treatment at the beginning of the study. In summary, our study demonstrates that former randomisation to 7.5 years of intensified insulin treatment, leading to a lowering of HbA 1c levels, retards the incidence of ischaemic foot ulcers in the long-term follow-up compared with randomisation to standard insulin treatment. The intensified-insulin-treated group also had a better skin microcirculation response compared with the standard-insulintreated group 5.5 years after the original SDIS ceased. Chronic hyperglycaemia negatively associated with endothelialdependent vasodilatation, which has been suggested as a culprit in the development of ischaemic foot ulcers. Thus, improving glycaemic control may help maintain better skin microcirculation and thereby decrease the risk of ischaemic foot ulceration and eventual leg amputation. Acknowledgements We dedicate this paper to Per Reichard (Department of Clinical Science and Education, Division of Internal Medicine, Karolinska Institutet, Stockholm, Sweden), who died in 2002 and to whom we are greatly indebted. Per was a pioneer in the treatment of type 1 diabetes, guiding patients and colleagues to improve diabetes care. In memory of Bertil Gazelius (Perimed, Järfälla, Sweden), who died in 1999, and Lars Erik Lindblad (Department of Clinical Science and Education, Section of Clinical Physiology, Karolinska Institutet, Stockholm, Sweden), who died in We thank C. Ihrman-Sandahl (Stockholm Heart Center, Stockholm, Sweden) and M. Pihl (Department of Diabetes and Metabolism, Karolinska Institutet, Karolinska University Hospital, Stockholm, Sweden) who performed the practical investigations. We thank L. Benson (Department of Clinical Science and Education, Karolinska Institutet, Stockholm, Sweden) for statistical advice. Funding This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. Duality of interest The authors declare that there is no duality of interest associated with this manuscript. Contribution statement KJ-U conducted the iontophoresis and, together with Per Reichard (PR), analysed the data and drafted an earlier manuscript that was never published due to illness of PR. Other co-writers BR and TN planned the follow-up part of this study, and analysed all the data from the iontophoresis investigation and the long-term follow-up of the present study, and drafted the manuscript. All authors took part in revising the manuscript and gave final approval of the version to be published. TN is responsible for the integrity of the work as a whole. References 1. Boulton AJ, Vileikyte L, Ragnarson-Tennvall G, Apelqvist J (2005) The global burden of diabetic foot disease. Lancet 366: Lehto S, Ronnemaa T, Pyorala K, Laakso M (1996) Risk factors predicting lower extremity amputations in patients with NIDDM. Diabetes Care 19: Brownlee M (2001) Biochemistry and molecular cell biology of diabetic complications. Nature 414: Feihl F, Liaudet L, Waeber B, Levy BI (2006) Hypertension: a disease of the microcirculation? Hypertension 48: Boulton AJ (2004) The diabetic foot: from art to science. The 18th Camillo Golgi lecture. Diabetologia 47: Veves A, Akbari CM, Primavera J et al (1998) Endothelial dysfunction and the expression of endothelial nitric oxide synthetase in diabetic neuropathy, vascular disease, and foot ulceration. Diabetes 47: (1993) The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. The Diabetes Control and Complications Trial Research Group. N Engl J Med 329: Reichard P, Nilsson BY, Rosenqvist U (1993) The effect of long-term intensified insulin treatment on the development of microvascular complications of diabetes mellitus. N Engl J Med 329: Nathan DM, Cleary PA, Backlund JY et al (2005) Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes. N Engl J Med 353: Cefalu WT (2005) Glycemic control and cardiovascular disease should we reassess clinical goals? N Engl J Med 353: Reichard P, Britz A, Cars I, Nilsson BY, Sobocinsky-Olsson B, Rosenqvist U (1988) The Stockholm Diabetes Intervention Study (SDIS): 18 months' results. Acta Med Scand 224: Gudbjornsdottir S, Cederholm J, Nilsson PM, Eliasson B, Steering Committee of the Swedish National Diabetes R (2003) The National Diabetes Register in Sweden: an implementation of the St. Vincent Declaration for Quality Improvement in Diabetes Care. Diabetes Care 26: Ludvigsson JF, Andersson E, Ekbom A et al (2011) External review and validation of the Swedish national inpatient register. BMC Public Health 11:450

8 14. Ragnarson Tennvall G, Apelqvist J, Eneroth M (2000) The inpatient care of patients with diabetes mellitus and foot ulcers. A validation study of the correspondence between medical records and the Swedish Inpatient Registry with the consequences for cost estimations. J Intern Med 248: Schaper NC, Andros G, Apelqvist J et al (2012) Diagnosis and treatment of peripheral arterial disease in diabetic patients with a foot ulcer. A progress report of the International Working Group on the Diabetic Foot. Diabetes Metab Res Rev 28(Suppl 1): Jensen-Urstad KJ, Reichard PG, Rosfors JS, Lindblad LE, Jensen-Urstad MT (1996) Early atherosclerosis is retarded by improved long-term blood glucose control in patients with IDDM. Diabetes 45: Holman RR, Paul SK, Bethel MA, Neil HA, Matthews DR (2008) Long-term follow-up after tight control of blood pressure in type 2 diabetes. N Engl J Med 359: Greenman RL, Panasyuk S, Wang X et al (2005) Early changes in the skin microcirculation and muscle metabolism of the diabetic foot. Lancet 366: Boyko EJ, Ahroni JH, Stensel V, Forsberg RC, Davignon DR, Smith DG (1999) A prospective study of risk factors for diabetic foot ulcer. The Seattle Diabetic Foot Study. Diabetes Care 22: Tuomilehto J, Borch-Johnsen K, Molarius A et al (1998) Incidence of cardiovascular disease in Type 1 (insulin-dependent) diabetic subjects with and without diabetic nephropathy in Finland. Diabetologia 41: Hamdy O, Abou-Elenin K, LoGerfo FW, Horton ES, Veves A (2001) Contribution of nerve-axon reflex-related vasodilation to the total skin vasodilation in diabetic patients with and without neuropathy. Diabetes Care 24: Petrofsky JS, McLellan K (2009) Galvanic skin resistance a marker for endothelial damage in diabetes. Diabetes Technol Ther 11: Ramsay JE, Ferrell WR, Greer IA, Sattar N (2002) Factors critical to iontophoretic assessment of vascular reactivity: implications for clinical studies of endothelial dysfunction. J Cardiovasc Pharmacol 39: Millet C, Roustit M, Blaise S, Cracowski JL (2012) Aging is associated with a diminished axon reflex response to local heating on the gaiter skin area. Microvasc Res 84: