Good health is more than the absence of disease, and the review by McNicholas and colleagues makes a strong argument that the companionship of pets contributes to wellbeing.1 People decide to keep pets for reasons that go far beyond health, and there are social benefits when animals promote conversation and friendship. The debate about health effects of pet ownership has focused on the major problems of cardiovascular disease, mental health, and allergy. A point to add to the review is that the benefits partly depend on the type of animal. In a prospective study of one year survival after a myocardial infarct, dog owners were more likely to survive than cat owners and people who did not own pets.2 Although exercise from dog walking might be a factor, the improved survival was related to social support independent of physiological status.

On the fringes of the debate about health and pets are some unusual observations about benefits and hazards of pets. The ability of some dogs to give an early warning of an epileptic fit or a hypoglycaemia attack is extraordinary.3 4 How this is done is uncertain, perhaps by an acute sense of smell or visual cues from the owner's behaviour and posture. Whether a dog that earns its living, by being an early warning system or guiding a blind person, is a “pet” is a question that can only be answered by the person with whom the animal lives: to be a pet, an animal should be valued for more than its utility.

On the opposite side of the debate are the infection hazards associated with pets (table). The hazards have been well described in a review by Guay.5 It is difficult to quantify the risk in relation to the popularity of pets, but the infections are either uncommon or uncommonly acquired from pets. The list could be lengthened to include cryptosporidioisis and Escherichia coli O157 acquired from lambs and calves if they are treated as pets on farms. Although the incidence may be low, the infections should not be forgotten when recommending pets.

In assessing the risks of infection, the health of the animal should be considered. The probability that a pneumonic illness in a budgerigar owner is due to psittacosis increases when the patient says that a pet bird has recently died. Puppies with diarrhoea are more likely to be a source of campylobacter than a healthy house trained adult dog. Many people will accept the risks of gastrointestinal infection in exchange for the pleasure of watching a young animal at play, and for the bond between pet and owner that grows with time. Risks can be reduced by awareness and hygiene, but adopting a house trained adult pet is an alternative. For an older person worried that a young pet's life expectancy would be much longer than their own, a mature dog or cat adopted from an animal sanctuary can be a happy, and perhaps a health promoting, companion.