We elucidated that following an emergency department attendance at another hospital (with a fall and minor head injury) the warfarin had been stopped partly due to “extensive facial bruising”.

Past medical history included mild hypothyroidism, essential hypertension, hip and knee replacements for osteoarthritis, and acne rosacea (diagnosed in 1985) for which she had been taking oral minocycline 100 mg modified release capsules once daily for 12 years. Additional medication included lansoprazole 30 mg once daily, paracetamol 1g four times daily, Adcal D3 (calcium/vitamin D3 supplement) three times daily, and a beclomethasone 100 μg inhaler.

On further questioning the patient reported that she had noticed a change in the pigmentation of the skin over several months and commented that friends and medical personnel had also remarked on facial “bruising” during this time.

On examination she was apyrexial with an irregularly irregular pulse rate of 80 beats/min, a blood pressure of 138/76 mm Hg, and a respiratory rate of 18/min. Atrial fibrillation was confirmed by electrocardiogram (ECG). The chest was clear to auscultation with oxygen saturation by pulse oximetry being 91% on air. Arterial blood gas measurements taken at the time confirmed this and demonstrated mild hypoxaemia with a pO₂ 8.9 kPa and normal pH and pCO₂. The abdomen was soft and non-tender and there were no focal neurological signs. There were no signs of deep vein thrombosis or pelvic masses.

Extensive pigmentation was present on the face and was described by the patient as similar on presentation as compared to when her warfarin was stopped. Symmetrical, bilateral, confluent, blue-grey macules were noted in a photosensitive distribution over the eyebrows, malar region, nasal bridge and chin (fig 1). Concomitant pigmentation of the sclera was also visible.

Additionally, the finding of blue-grey discolouration of the proximal portion of the nails bilaterally in the hands further supported the diagnosis of minocycline induced hyperpigmentation (fig 2).

A chest x-ray was normal; however, the suspected diagnosis of a pulmonary embolus was confirmed by computed tomography pulmonary angiography (CTPA) which indicated filling defects within the left upper lobe pulmonary tree (within the proximal apicoposterior and lingua pulmonary arteries). There was also a degree of subsegmental atelectasis in the right lower lobe and mild cardiomegaly with small pleural effusions (fig 3).

While each patient should be reviewed on their own merits, the decision to stop anticoagulation in this group of patients is less well supported by evidence. The actual risk–benefit ratio in continuing treatment appears to favour the beneficial end of the spectrum, and physician’s unfounded or misaligned concerns are demonstrated by systematic review of the literature.^2,3

Patients with a high risk of falling do have a higher incidence of intracranial bleeding than those who are at low risk of falls. As we would expect, the addition of warfarin does result in a higher mortality from intracranial bleeding, but interestingly it has a minimal effect on the occurrence rate of such.^4,5 Furthermore, although the risk of falls rises with age this does not appear to negate the benefits of warfarinisation; the BAFTA trial⁵ demonstrated an annual risk of stroke of 1.8% in the warfarin arm versus 3.8% in the aspirin arm and an absolute reduction of 2% in patients over 75 years of age.

Importantly, in those with a high validated clinical prediction rule CHADS2 score (2 and above) the risk of embolic stroke has been shown to far outstrip that of bleeding;⁶ our patient had a score of 2 which would put her annual risk of stroke at approximately 4–6%. Warfarin has been shown to reduce the composite mortality of stroke, intracranial haemorrhage, myocardial infarction and death in this group.⁷ This means that despite the increased mortality from traumatic head injury on warfarin following a fall, those who are at high risk of embolic stroke confer benefit from anticoagulation due to their even greater embolic stroke risk.

Minocycline induced skin pigmentation is a well recognised drug side effect that usually begins at around 3 months but has a prevalence as high as 41% of patients at 1 year in certain patient groups.¹ There are thre main recognised types.⁸ Type 1 involves pigmentation of old scarring, especially acne scarring on the face; type 2 shows pigmentation of normal skin on the lower legs and forearms; and type 3 favours a more photosensitive distribution.⁸ In addition, pigmentation of other structures such as teeth, nails, bones, sclera and the thyroid are recognised.^9–11 Pigmentation types 1 and 2 are associated with extracellular iron and melanin deposition in macrophages, whereas type 3 patterns are due to basal keratinocytes melanin accumulation only.⁸ The discolouration is generally reversible upon discontinuing medication but can be permanent, especially with the photosensitive type.⁸

Pigmentation of the skin can be mistaken for bruising¹² and, in this case, we were provided with a useful highlight of a complex clinical decision that can be aided by an evidenced based approach.