Role of nicotinic acid and nicotinamide in nicorandil‐induced ulcerations: from hypothesis to demonstration

Abstract

Nicorandil, a nicotinamide ester, was first reported to be involved in the induction of oral ulcers in 1997. Since then, many reports of single or multiple nicorandil‐induced ulcerations (NIUs) have been reported. We hypothesised that in the case of high‐dosage nicorandil or after an increased dosage of nicorandil, nicotinic acid and nicotinamide (two main metabolites of nicorandil) cannot appropriately merge into the endogenous pool of nicotinamide adenine dinucleotide/phosphate, which leads to abnormal distribution of these metabolites in the body. In recent or maintained trauma, nicotinamide increases blood flow at the edge of the raw area, inducing epithelial proliferation, while nicotinic acid ulcerates this epithelial formation, ultimately flooding the entire scar. We demonstrate, by comparison to a control patient non‐exposed to nicorandil, an abnormal amount of nicotinic acid (×38) and nicotinamide (×11) in the ulcerated area in a patient with NIUs. All practitioners, especially geriatricians, dermatologists and surgeons, must be aware of these serious and insidious side effects of nicorandil. It is critical to rapidly reassess the risk–benefit ratio of this drug for any patient, and not only for those with diverticular diseases.

Introduction

Nicorandil (ADANCOR^®, IKOREL^®, etc.), a nicotinamide ester [N‐(2 hydroxyethyl) nicotinamide nitrate ester], is the first and only association of nitrates and potassium channel activators (PCA, (Figure 1). Nicorandil opens adenosine triphosphate (ATP)‐sensitive K+ channels, causing dilatation of peripheral and coronary resistant arterioles. Nicorandil also has a nitrate moiety that dilates systemic veins and epicardial coronary arteries. The nicotinamide moiety that supports the PCA' effect is present in its metabolites 1, 2. Nicorandil is a second or third line medical treatment of stable angina. The recommended oral doses range from 10 to 20 mg, twice daily.

Figure 1.

Chemical plane structures for nicorandil, nicotinamide and nicotinic acid.

Nicorandil goes through extensive hepatic metabolism; its metabolites are almost eliminated in urine. Approximately 1% of a dose is excreted unchanged in urine and faecal excretion accounts for less than 2%. The two main biotransformation pathways are denitration leading to pharmacologically inactive N‐(2‐hydroxyethyl) nicotinamide and reduction of the alkyl chain leading to nicotinamide/nicotinic acid 3.

Nicotinamide and nicotinic acid (Figure 1), as well as their derivatives: N‐methylnicotinamide, 2‐pyridone, 4‐pyridone, nicotinamide oxide, etc., merge into the endogenous pool of nicotinamide adenine dinucleotide/phosphate (NAD/NADP) 4.

This merging, which is known to be saturable (from nicorandil metabolism) may contribute to a small, yet significant accumulation of nicorandil and N‐(2‐hydroxyethyl) nicotinamide after repeated administration 3, 4.

Nicorandil was first found to be involved in the induction of oral ulcers by Reichert et al. 5. Five years later, Watson et al. 6 described anal ulcers. Since then, many reports of single or multiple nicorandil‐induced ulcerations (NIUs) in oral, anal, perianal, vulvovaginal, perivulval, penile, gastrointestinal, colonic, peristomal, skin and ocular locations have been reported 7, 8, 9, 10, 11, 12, 13, 14, 15. Fistulae are now being investigated 16.

Until now, PCA and nitric oxide (NO) dilatators were not known to induce ulcerations by themselves, whereas drug‐induced ulcerations have been recognised as a side effect of nicorandil.

Characteristics of NIUs

NIUs are painful, large, deep, well‐circumscribed and clean, without oedema, non‐healing, progressively worsening, histologically non‐specific, and can lead to local fistulae or spread of fistulae into adjacent organs. Surgical cleaning is ineffective and can exacerbate morbidity. Ulcers and fistulae heal only after drug withdrawal. The delay expected for wound healing is 15 days up to 4 months. NIUs mostly occur at vulnerable sites, being peristomal after any procedure that creates a stoma, penile after circumcision and perianal after haemorrhoidectomy. Under cardiologists' aegis, nicorandil replacement by a classic NO donor can be done without complication 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16.

Hypotheses

Since 1997, we have been trying to elucidate the pathogenesis of NIUs. Different hypotheses have been proposed: the excretion of a toxic metabolite, a vascular access steal process or a defect in actin filament contraction which may inhibit the early inflammatory phase of wound healing through potassium channel activity. Patel and Harding 17 demonstrated that the first two hypotheses were unconvincing when considering both the ulcer location as well as the findings that ‘the sites were typically well‐vascularised, the surrounding skin looks healthy and histological support was lacking’.

In 2006, we suggested ‘NIUs on the mucosa as well as on fragile skin was related to previous physical aggression or occurred where some form of cutaneous pathology existed before the ulceration’ 18.

In 2008, we further hypothesised that ‘during prolonged high‐dose treatment with nicorandil, or after increased dosage, nicotinic acid may accumulate outside the endogenous pool of NAD/NADP, and become abnormally distributed’. This event could then promote ‘the genesis of ulcerations in a targeted compartment, such as mucosa or skin rendered vulnerable because of physical aggression or cutaneous pathology’ 19.

In 2012, we reinforced our hypothesis with the notion that in the case of ‘external ulceration’, the nicorandil ulcer is typically localised to the area where the trauma occurred 20.

Until this time, our hypothesis was supported for nicotinic acid, on an in vivo acid with 4·9 pKa and for nicotinamide on a tumour radiosensitizer increasing blood perfusion 21.

Minoxidil, another PCA, provided us a new pharmacological rationale. In 1995, Brenda et al. 22 compared four topical drugs including minoxidil on their ability to promote the healing of open wounds in rats. The closure of the first lesion occurred on the 49th post‐operative day and Brenda noted that wounds in the minoxidil group showed a small amount of inflammatory infiltrate only under the raw area and an extensive band of new epithelium adhering to the subjacent granulation tissue was present. Brenda stated that ‘‘the mechanism whereby minoxidil increases epithelial proliferation may therefore be related to increased blood flow in the tissues at the edge of the raw area’’.

Thus, our hypothesis of a ‘chemical ulceration’ induced by nicorandil was that, in the case of high‐dosage or after an increased dosage of nicorandil, nicotinic acid and nicotinamide derivatives (two main metabolites of nicorandil) may be unable to merge into the endogenous pool of NAD/NADP, leading to abnormal in situ distribution of these molecules in the body. Additionally, in the case of recent or maintained trauma, nicotinamide increases the blood flow at the edge of the raw area, inducing epithelial proliferation, while nicotinic acid ulcerates the new epithelial formation, ultimately flooding the whole scar.

The patient

In October 2012, an 84‐year‐old woman treated with acenoumarol, methotrexate, potassium chloride, ivabradine, transdermal nitroglycerine and nicorandil (20 mg/day for 15 months then 40 mg/day for 7 months until hospitalisation) was admitted to the department of dermatology with two painful ulcers, one localised on the right external malleolus (Figure 2) and one on the right front ankle (Figure 3). Both failed to scar by 6 months. NIUs were diagnosed and treatment with nicorandil was discontinued. With informed consent of the patient, punch‐biopsies (Ø 4‐mm) were made the day treatment ended: three in the centre of the right external malleolus ulcer and three on healthy but painful skin, 10 cm above the ulcer site. Simultaneously, three similar punch‐biopsies in the corresponding skin location were performed on a control patient non‐exposed to nicorandil (one of the authors). The patient's condition improved in 2 weeks and complete healing was achieved in 9 weeks.

Figure 2.

Right external malleolus ulcer induced by nicorandil.

Figure 3.

Right front ankle ulcer induced by nicorandil.

Procedures and analytical conditions

Skin biopsies were pooled and snap‐frozen in liquid nitrogen before being cut into 10‐µm thick sections with a cryostat. After freeze‐drying and precise weighing (±0·01 mg), dried tissues were suspended in 5 ml of molecular biology grade distilled water, and vortexed for 5 min. After centrifugation (5 min at 1000 g), 4 ml of supernatant was extracted by mechanical agitation with Varian Toxi‐Tube^® A (Agilent‐Technologies.Fr, Massy, France) in the presence of 20 µl of cyproheptadine (1 mg/ml) used as internal standard (IS). After centrifugation, the upper organic layers were transferred into glass tubes and evaporated under a gentle stream of nitrogen. Silylation was performed by incubating the dried residues with 30 µl of NO‐Bis(trimethylsilyl)trifluoroacetamide (BSTFA) for 35 min at 80°C. After the samples cooled and a final centrifugation, 1 µl of each silylated sample was injected into the chromatographic system.

DSQII MS apparatus (Thermo‐Fisher‐Scientific.Fr, Illkirch, France) was coupled with a Focus GC system equipped with a DB5MS column (Agilent‐Technologies.Fr) of 30 m length, 0·25 mm diameter and 0·25 µm of film thickness. The oven temperature was maintained at 70°C for 1 min, ramped to 230°C at a rate of 25°C/min, then up to 300°C at a rate of 10°C/min, and finally plateaued at 300°C for 0·5 min. The mass spectrometer was operated in the electron impact mode with 70 eV energy.

Samples were acquired in the selected ion monitoring (SIM) mode. The selected m/z values were 106, 136 and 180 for nicotinic acid, 51, 75 and 179 for nicotinamide and 96, 215 and 287 for cyproheptadine (IS). Data were processed with XCalibur^® software (Thermo‐Fisher‐Scientific.Fr) and estimation was performed relative to IS amount using freshly prepared standard aqueous solutions ranging from 1 to 100 µg/ml of nicotinic acid or nicotinamide (Sigma‐Aldrich.Fr, Saint‐Quentin Fallavier, France). Final sample concentrations were expressed as micrograms of nicotinic acid or nicotinamide per milligram of dried skin tissue.

Results

Dosages summarised in Table 1 are in accordance with our hypothesis involving nicotinic acid and nicotinamide in the induction of NIUs.

Table 1.

Analytical data obtained on skin biopsies

Sample	Nicorandil‐free volunteer	Nicorandil‐treated patient
		Non‐ulcerated area	Ulcerated area
Nicotinic acid (µg/mg)a	≤1	21	38
Nicotinamide (µg/mg)a	≤1	4	11

^aData expressed as micrograms per mg of dried skin biopsy.

Compared to a non‐exposed control, the patient with NIUs had an abnormal elevated amount of i‐nicotinic acid in ulcerated and non‐ulcerated areas (×38 and ×21) and ii‐nicotinamide (×11 and ×4). Moreover, there was an accumulation of the two metabolites in ulcerated area and in non‐ulcerated area, respectively with 3·45‐fold and 5·25‐fold increases in nicotinic acid.

Conclusion

In March 2012, the results of a French survey led to a ‘health care professional letter’ for the recognition and management of this adverse drug reaction. NIUs mostly occur in the elderly and fragile population with a median age of 80·6 years (39–94). A majority of NIUs have been reported to occur after high‐dosage (≥40 mg/day) or after an increased dosage of nicorandil 23.

NIUs are totally underestimated. Smith et al. 24 reported that ceasing the administration of nicorandil led to an improvement in the skin ulcerations of 18 of 21 patients with either perianal or parastomal ulcerations. A previously injured area, whether recent or maintained, symptomatic or not, is required for development of NIUs

NIUs are easy to treat. Simply withdrawing nicorandil treatment results in complete healing in 15 days up to 4 months, mainly depending on the location of the ulcer. Surgical cleaning is not only ineffective but can exacerbate morbidity by digging a furrow into the tissue for excess nicotinic acid . Under the cardiologists' aegis, nicorandil replacement by a classic NO donor can be carried out without complication.

All practitioners but especially geriatricians, dermatologists and surgeons must be aware of these serious and insidious side effects of nicorandil. It is critical to rapidly reassess the risk–benefit ratio of this drug for any patient, and not only for those with diverticular diseases 24, 25.