Active post-marketing surveillance of the intralesional administration of human recombinant epidermal growth factor in diabetic foot ulcers

Background After several exploratory and confirmatory clinical trials, the intralesional administration of human recombinant epidermal growth factor (hrEGF) has been approved for the treatment of advanced diabetic foot ulcers (DFU). The aim of this work was to evaluate the effectiveness and safety of this procedure in medical practice. Methods A prospective, post-marketing active pharmacosurveillance was conducted in 41 hospitals and 19 primary care polyclinics. Patients with DFU received hrEGF, 25 or 75 μg, intralesionally 3 times per week until complete granulation of the ulcer or 8 weeks maximum, adjuvant to standard wound care. Outcomes measured were complete granulation, amputations, and adverse events (AE) during treatment; complete lesion re-epithelization and relapses in follow-up (median: 1.2; maximum 4.2 years). Results The study included 1788 patients with 1835 DFU (81% Wagner’s grades 3 or 4; 43% ischemic) treated from May 2007 to April 2010. Complete granulation was observed in 76% of the ulcers in 5 weeks (median). Ulcer non-ischemic etiology (OR: 3.6; 95% CI: 2.8-4.7) and age (1.02; 1.01-1.03, for each younger year) were the main variables with influence on this outcome. During treatment, 220 (12%) amputations (171 major) were required in 214 patients, mostly in ischemic or Wagner’s grade 3 to 5 ulcers. Re-epithelization was documented in 61% of the 1659 followed-up cases; 5% relapsed per year. AE (4171) were reported in 47% of the subjects. Mild or moderate local pain and burning sensation, shivering and chills, were 87% of the events. Serious events, not related to treatment, occurred in 1.7% of the patients. Conclusions The favorable benefit/risk balance, confirms the beneficial clinical profile of intralesional hrEGF in the treatment of DFUs.


Background
Pharmacovigilance and post-marketing studies are key elements to monitor the safety and effectiveness of approved drugs. They are excellent scenarios to confirm and extend the safety profile and efficacy data acquired in previous clinical trials.
Products for the treatment of diabetic foot ulcers (DFU) available in the market have been scarcely effective in advanced and/or ischemic lesions [1]. Necrotic tissue, sepsis, inflammation and wound proteases impair the effective distribution of the active pharmaceutical ingredients when they are administered in topically [2].
A new procedure has been recently developed for the treatment of these advanced DFU based on the intralesional infiltration of recombinant, human epidermal growth factor (rhEGF) as a lyophilized formulation under the brand name Heberprot-P®. The rationale of this procedure has been published and reviewed recently [2,3].
Exploratory [4][5][6] and confirmatory [7] clinical trials have been successfully completed with this product in patients suffering from advanced DFUs with potential amputation risk (reviewed in ref. 3). These trials demonstrated that the intralesional administration of rhEGF accelerates the wound healing process. The generation of useful granulation tissue on the ulcer bed facilitated its closure by either second intention or a skin graft. The product's safety profile was acceptable and the benefitrisk analysis of its use yielded a largely favorable balance. These results granted the approval by the Cuban Regulatory Authority (CECMED).
In April 2007 this treatment was included in the country's Basic Drugs List and its use was extended to all Angiology services, firstly in the secondary healthcare level (hospitals) and subsequently in the primary level (polyclinics).
At the same time, postmarketing active surveillance was initiated in order to evaluate the effectiveness and safety of the drug in the current medical practice. This article reports the first results of this surveillance.

Design and participants
A multicenter, prospective, intensive, post-marketing surveillance was conducted on patients treated from May 2007 to April 2010; follow-up extended to December 2011. The study was coordinated by the Center for the Development of Pharmacoepidemiology (CDF) from the Cuban Ministry of Health and its national network. There were 41 participating hospitals from the 15 Cuban provinces and 19 polyclinics from 7 provinces.
Patients, more than 18 years-old, suffering from 1 cm 2 or larger DFUs were prescribed with rhEGF and included. The indication comprised Wagner's [8] grades 3 and 4 DFU, but some patients with grades 1, 2, and 5 were included, under off-label use according to the physicians' decision. Patients with diabetic coma, uncontrolled heart, renal or liver disease, history or suspicion of malignancy, pregnancy or breastfeeding were not treated. Prescription was part of the patient's regular care, not induced by the present work. The protocol was approved by the CDF Institutional Review Board. The confidentiality of the patients' personal data was preserved.
Lesion etiology was classified as ischemic or not. When the ankle/brachial index (ABI) was available, patients with values below 0.75 were considered ischemic, following the same criterion used in the clinical trials. Otherwise, clinical signs of ischemia such as absence of pulses (femoral-popliteal, tibial or distal), intermittent claudication, pain, local atrophy, coldness, and hair loss were taken into account for this classification.
Since it was not feasible to standardize the ulcer size measurement among so many clinical sites, lesion extension-location was considered in three categories: (i) simple, if the ulcer covered only one region of the foot (toes, dorsum, sole, internal edge, external edge, except calcaneus); (ii) complex, if the extension of the ulcer comprised several regions but not the heel, and (iii) calcaneal if this region was involved.

Intervention
The treatment was as in-patients, although ambulatory care was allowed if the subject could attend the treatment visits. The standard care included the patient's metabolic control, lesion area sanitation and systematic cures, sharp debridement of the necrotic or infected tissue with minor amputation of the affected zones if necessary, and moist gauze dressing. Wide spectrum antimicrobial drugs were prescribed in patients exhibiting local clinical signs of infection. Pressure off-loading of the affected zones was recommended as well.
The product (Heberprot-P®, Heber Biotec, Havana) is lyophilized, containing 75 or 25 μg of rhEGF per vial, to be dissolved with 5 ml of water for injection. In every visit this volume was distributed throughout the lesion in 0.5-1 ml injections. The solution was injected first into the dermo-epidermal junction at equidistant points all over the lesion contours and then downward into the wound bottom to ensure a uniform distribution. The needle was changed for each puncture. The product label indicates injections 3 times per week (tpw) on alternate days. However, in some cases the physicians decided to modify the schedule to daily, twice or once per week. A treatment cycle continued until complete granulation was achieved, lesion closed by autografting or 8 weeks maximum.
The rhEGF dose (25 μg or 75 μg) was selected according to the label (25 μg in ulcers smaller than 20 cm 2 and nonischemic), but the choice was also determined by each physician's criteria, his/her experience with this product, and the availability of either drug presentation in the healthcare unit at a certain moment.

Outcomes and measurement methods
The main effectiveness variable was lesion complete granulation, evaluated by direct visual inspection. It was defined as productive material, able to mediate the complete lesion closure by second intention or an autologous skin graft. Macroscopically, it was characterized by the presence of reddish, diffused, dispersed and lustrous miliary granular formations that bleed easily after manipulation.
Secondary outcomes were time-to-complete granulation, need for amputation and its type during treatment. During the follow-up period wound closure (complete re-epithelization), relapses, and patient's survival were evaluated.
The variables used to evaluate safety were the adverse events (AE) during the treatment period, considering the type of event, organ and system affected [9], its seriousness and causality relation with the treatment. The latter was done according to the World Health Organization (WHO) algorithm [10] only for serious adverse events (SAE) by the provincial pharmaco-epidemiologists and a consulting multidisciplinary team created ad hoc for this purpose. SAE reported were included in the Cuban Pharmacosurveillance System Database, which guarantees the absence of duplicated data when processing reports received from different sources.
Information was gathered by the physicians in case report forms (CRF) which covered the in-patient treatment period. Patients' final outcome was collected in a further follow-up visit, done by the hospital or municipality pharmacoepidemiology staff, scheduled annually. Emphasis was placed on the training of the personnel in all aspects related to the drug safety evaluation.
Deaths during treatment were further investigated by review of the clinical records and, if necropsy was performed, its report. Causes of death were taken from the death certificates, coded according to the International Classification of Disease-10.
The occurrence of any type of cancer was actively investigated through: (i) direct interview to the patient in the follow-up visits; (ii) cross-search in the National Cancer Registry (NCR), or (iii) findings in the National Mortality Registry. The latter cross search was also useful for the identification and confirmation of follow-up period deceases.

Statistical analyses
One patient could be treated more than once for the same or a different lesion at different moments. Then the experimental unit considered was the treatment cycle, since each of them generated one CRF. Short term AE were evaluated on this basis. The lesion was the unit taken into account for the analyses of granulation, amputations, healing, and relapses. Survival and long term AE were done on patient basis.
The statistical treatment of the data was conducted with the PASW 18.0 software. Measures of central tendency and dispersion such as mean, median, standard deviation (SD), 95% CI, quartile range (QR), minimal and maximal values, were calculated to describe quantitative variables. Graphical normality analysis (QQPlot) and Kolmogorov-Smirnov test for goodness of fit were applied. The chisquare values were calculated to evaluate dependency among qualitative variables. Time variables were estimated using Kaplan-Meier plots and compared with the log-rank test. Logistic or Cox regression models were adjusted to assess the influence of control variables on the outcomes. For the variable-association analyses only the results from the first treatment cycle on each patient were taken into account to avoid considering non-independent observations. Safety related variables were subjected to descriptive statistics. A Bayesian approach was used for the benefitrisk ratio analysis.

Characteristics of the subjects
CRFs were available from 1788 subjects bearing 1835 DFU. This population represents approximately 80% of all the DFU patients treated with rhEGF in the study period, according to weekly reports received at CDF. Of them, 1676 (93.7%), were seen in hospitals and 112 (6.3%) in polyclinics. Most of the patients (1729; 96.7%) received one treatment cycle; 56 (3.1%) were given two cycles; four of them simultaneously on two different lesions and 3 (0.2%) had three cycles. Therefore the whole patient population received 1851 treatment cycles with rhEGF.

Treatment compliance
There were some treatment schedule deviations according to doctors' decisions: 267 Wagner's grade 1, 2, or 5 ulcers (15% of all lesions) were treated as off-label indications; 281 ischemic lesions (36%) were treated with the 25 μg dose; 50 lesions were treated with both doses, considered as 75 μg for the analyses; the three tpw schedule was not followed strictly (see next paragraph); 48 treatment cycles comprised more than 24 injections.

Amputations
At the end of treatment 214 patients required 220 amputations (12%) ( Table 2); 23 of them were disarticulations, 26 transmetatarsal and 171 major amputations (9.3% of the lesions). Ischemic ulcers caused more amputations. Most of the amputations (85%) were in cases with Wagner's 3-5 ulcers. Calcaneus location of the ulcers was also an unfavorable factor for the amputation outcome (18.5% vs. 11.0% in non-calcaneus ulcers).

Adverse events
A total of 4171 AE were reported (70 different types) in 46% (856/1851) of the treatment cycles in 838 subjects (47%). The AE occurring in more than 1% of the patients and the maximal number of repetitions of the event in any subject are summarized in Table 3. Pain and burning sensation at the administration site, shivering and chills account for 87% of all the AE reported. Their frequency decreased as the treatment continued: from more than 14% in the first application to less than 2% after 40 days. More than 85% of the events were mild or moderate and easily manageable. Except for local infection, all common AE were more frequent with the 75 μg dose. The rate of local infection as an AE (4%) was not dependent on the baseline infection status of the ulcer ( Table 4).
The 31 SAE reported are shown in Table 5. Two of them (ketoacidosis and gastroenterocolitis) were unrelated to the treatment. The other 29 were classified as conditional or possible, since there was a temporal relation with treatment but there are alternative explanations to the clinical findings. Local infection accounted for 32% of the SAE and cardiovascular syndromes for 14 SAE. Thirteen of the latter patients had antecedent of cardiovascular disease (hypertension, ischemic cardiopathy, or arrhythmia).

Follow-up: wound closure
Post-treatment follow-up information was obtained from 1620 subjects (91%) with 1659 ulcers, at least once during the 4 years after the end of treatment. The follow-up period (median: 1.2; maximum 4.2 years) comprised 2270.6 years-person. The baseline and demographic characteristics of the followed-up subjects do not differ from the whole patient population (data not shown). Complete ulcer re-epithelization was achieved in 55% of all the lesions ("intention-to-treat" evaluation), 61% of those evaluated (Table 2). In a logistic regression model significant positive influence on this outcome was shown by: non-ischemic etiology (OR: 2.1; 95% CI: 1.6-2.6), age (1.012; 1.002-1.022, for each younger year), and nothistory of amputation before treatment (1.7; 1.3-2.1). Relapses occurred in 5% person-years, regardless of the ulcer etiology or other characteristics (Table 2).

Survival and long-term safety
During treatment or follow-up 352 patients (20%) died.  Figure 2 illustrates the effects of healing, etiology, and amputation on survival through the corresponding Kaplan-Meier plots.

Discussion
The postmarketing surveillance covered all the Angiology wards that manage DFU in Cuba and some primary care units that have incorporated this drug as part of the diabetic patients integral care program. The demographic and baseline characteristics of the subjects parallel those of people with DFU that took part in the clinical trials with this procedure [3][4][5][6][7]: predominantly diabetes type 2, median age 65 years, approximately equal gender distribution, high proportions of hypertension, ischemic cardiopathy, and previous ulcers, as well as the ethnic distribution   of the Cuban population [11]. Co-morbidities such as hypercholesterolemia, arterial hypertension and ischemic cardiopathy have been reported in other settings [12] too. Nearly half of the subjects had ischemic ulcers, mostly Wagner's grades 3-4, which singularizes the application of intralesional hrEGF from other procedures that are mainly indicated in less advanced lesions [13]. Calcaneus ulcers were not excluded, which is also an unfavorable condition [14]. Some deviations from the labeled pattern of hrEGF application were seen, which were due to doctors' preference, according to their judgment of the patient's evolution. In some cases, ambulatory patients could not attend all the visits. The dose deviations were also determined, eventually, by product availability at a given moment. The off-label use of medicaments, particularly biologics, is not a seldom practice in many settings as part of the common medical practice [15,16].
Treatment interruptions were mainly due to failure that determined amputation mostly in ischemic, Wagner's 3-5 lesions. Very few were because of patients' intolerance (AE or voluntary abandonments). Hypergranulation is noteworthy since it is due to the same mechanism of healing enhancement. It was easily controlled by tissue excess removal and there is no information of related sequels, such as hypertrophic scars or keloids.
The results of the previous clinical trials with intralesional hrEGF (79.4% granulation in 296 patients; 95% CI: 74.8-84.0%) were confirmed in this study. The reepithelization rate found falls within the pooled analysis of the clinical trials too: 59.2%; 95% CI: 53.4-64.5% [3]. Granulation response showed good predictive value for final healing in this series [17]. The relapse rate (5% yearpersons) is better than reported [18,19]. For example, Boulton et al. [20] refer an estimate of 50-70% for the UK and 34% after one year follow-up in other series. The fact that postmarketing data confirm the results of the clinical trials is an important feature that endorses the use of intralesional rhEGF in medical practice for the treatment of advanced DFU. Topical EGF [21,22] or other growth factors [23,24] have been reported in smaller and/or less advanced ulcers. However, there is scarce information on their effectiveness in regular medical practice or have not fulfilled the expectations raised from clinical trials [25].
The main variables with influence on the treatment effectiveness, both for granulation and re-epithelization, were ischemia and age, which are well-known adverse prognosis factor for ulcer healing [26]. However, the granulation and healing rates attained indicate that the effect of the product is evident in ischemic patients too. Ischemia was considered when ABI < 0.75 or clinically. The ABI cutoff for ischemia varies among different guidelines, from 0.7 [27,28] to 0.9 [29]. Consequently, 40 subjects (85% granulation; 80% re-epithelization) classified as nonischemic in this work would fall into a "mild" ischemia category (with scarce clinical expose) according to other thresholds. This supports the notion that the product is also effective in patients with impaired blood supply.
Although this work did not include a non-treated control group to measure the impact of the treatment on amputation rate, other independent investigations in additional  representing a summary of the evidence provided by the data in favor of benefit (red), as opposed to risk (blue). A value larger than 1 means a favorable benefit-risk ratio. In this case: Bayes factor = 5.4; difference between probabilities: 61% (95% CI: 59%-64%). patient series have made such comparisons. González-Acosta et al. found a major amputation rate reduction from 26.7% to 8.3% when intralesional rhEGF was added to conventional standard care [30]. García-Herrera et al. report a 43.1% to 8.1% major amputation rate reduction [31]. These data are consistent with the 9.2% major amputation rate found in the present work.
The most frequently observed AE such as pain and burning sensation at the injection site, shivering, chills, and local infection coincide with the reported intralesional hrEGF safety profile [3][4][5][6][7].
Local infection is one of the most fearsome AE because it frequently leads to amputation. Approximately 60% of amputations are preceded by infected ulcers [32]. Local infection was classified with "possible" causality relationship with the treatment. This is difficult to evaluate since most of these ulcers were infected at baseline. This diagnosis was defined by signs such as edema, ulcer border redness, and secretion. Heberprot P® labeling says that these signs should be cleared (antibiotics, debridement) before the product is applied. Even though, a negative culture is never obtained. It is unlikely that the injection procedure contributes to spread any subclinical infection as shown by the fact that the occurrence of this AE was not associated to the baseline infection of the ulcer.
There is not enough information to relate cardiovascular AE and deaths to the treatment, since the patients have frequent antecedents of cardiovascular diseases. The relationship of diabetes and cardiovascular diseases is well known, moreover if DFU is present [33,34]. In data from the national death certificate database and the Annual Health Report [35] for 2007-2010, the death rates due to these causes in people with diabetes among the three main causes of death in the certificate double that of the general population. This death profile agrees with the literature for DFU patients [36,37].
A particular concern on the use of growth factors is the possibility of development or stimulation of a pre- existing malignancy. Contrary to PDGF, EGF cannot initiate malignant transformation [38] and results with EGF in experimental models have not demonstrated tumor promotion consistently [3,39]. Diabetes is a known cancer risk boosting condition [40][41][42][43]. The data do not support any cancer promotion by rhEGF treatment.
Giovannucci et al. discuss the possible mechanisms of the diabetes-cancer association, either by common risk factors or specific mechanisms in diabetes metabolism and management [44]. This aspect has to be further investigated and followed in the product risk-management program.
A B C Figure 2 Kaplan-Meier curves for survival time according to A) whether the ulcer healed or not; B) ulcer etiology or C) whether the affected limb was amputated or not.

Conclusions
In summary this postmarketing, prospective study is consistent with earlier findings from exploratory and confirmatory clinical trials, while extending the data to a larger clinical setting. The positive findings, related to the clinical evolution of the patients treated at secondary and primary Cuban healthcare levels, suggest a relevant role of this drug in the treatment of DFUs. Further clinical research and post-marketing information from other countries should enrich the evidence shown in this paper.