As I approach the threshold of old age, I wonder whether I will accept the process of crumbling with good grace, or discover that I have a strong desire to postpone death. I suspect that like most people I won’t be consistent and won’t want to discuss the matter. This is a dilemma raised by this study, one of several dozen already this year to bear the name of Harlan Krumholz. He looks at survival following a diagnosis of colorectal cancer in people aged 67 and over: if they have three or more comorbidities, their life expectancy is so poor that it is kinder not to go screening for bowel cancer.

Inhaled insulin has arrived, but is it safe and effective? The answer which emerges from this meta-analysis of the main studies is that we simply don’t know: there are no data going beyond six months, which is a short time in the life of a diabetic. A mild cough and a small reduction in lung function may be fine for a few months, but what about, say, lung fibrosis over several years, or in a decade or two? All we know is that short acting insulin can be used as effectively by this route as subcutaneously in the short term.

Many adults with asthma get by on a puff or two of salbutamol every now and again, but for the rest, adding a regular inhaled corticosteroid has been standard practice for over 30 years. The only major advance in that time has been the introduction of long-acting inhaled β-adrenergic agonists such as salmeterol, almost always in patients already using steroid inhalers. So when a systematic review appeared recently (Ann Intern Med 2006;144:904) suggesting that salmeterol worsens outcomes in asthma, I grew slightly alarmed. But this short rejoinder from the Canadian Asthma Group criticises that conclusion by pointing out the limitations of the review, especially its omission of the vast majority of patients who inhale steroid alongside their salmeterol. They advise continuing present prescribing and doing a new meta-analysis of the studies which used both agents.