Dr Najim highlights the valuable resource of the UK Prescribing Observatory for Mental Health (POMH) which appears to show that National Health Service trusts record suboptimal levels of metabolic monitoring and, indeed, of physical examination of high-risk patients prescribed antipsychotic medication. We would be most interested to know whether the POMH database can help highlight monitoring rates in those taking antipsychotics for indications other than schizophrenia, particularly bipolar disorder and dementia. Further, are there data on metabolic monitoring in individuals taking depot antipsychotic medication? This has been a question very much overlooked in the literature to date.

Dr Reed rightly queries whether the recommendation to screen for cardiometabolic problems is evidence based. He is no doubt aware of the controversy regarding screening for depression and for dementia when screening is not necessarily translated into measurable patient benefit. We would argue that the case for screening for cardiometabolic risk has strong face validity and at least a moderate evidence base that does justify our recommendations. We concede, however, that the detail of how much and how often is not fully resolved and is disputed in the current guidelines. The case for cardiometabolic monitoring is supported by the undeniably large prevalence of the problem. Some studies suggest that as many as 90% of patients with chronic schizophrenia maintained on antipsychotics have at least one clinically important cardiometabolic risk factor. ^{Reference Bell, Farmer, Ries and Srebnik1} Further, in this population, the risk is at least in part iatrogenic, thereby promoting the responsibility of the medical profession to detect and deal with it. Direct evidence comes from guideline implementation studies. Screening guidelines do seem to increase monitoring rates, although the increase is less than is often hoped. We recently examined this using a meta-analysis of screening rates before and after guideline implementation. ^{Reference Mitchell, Delaffon, Vancampfort, Correll and De Hert2} Seven studies have directly monitored rates in the same sample before and after guideline introduction and these reported on glucose surveillance. These studies showed a significant 15.4% (95% CI 4.8–25.9) increase (χ² = 8.1; P = 0.005) in glucose testing following the introduction of guidelines. This increase is significant but nevertheless rather disappointing, although when combined with gradually increasing awareness of metabolic complications could increase further with time.

Another type of evidence is the additional yield of significant complications found after the introduction of a systematic screening or surveillance programme. Several such studies exist but, as far as we are aware, none have randomised a group to metabolic screening and no metabolic screening, for ethical reasons. In non-randomised studies the yield from systematic monitoring for cardiometabolic problems is appreciable. For example, Kusumi et al ^{Reference Kusumi, Ito, Honda, Hayashishita, Uemura and Hashimoto3} began testing 537 patients who had schizophrenia but no pre-existing diabetes in June 2008 across 25 Japanese hospitals. At baseline, only 51% had a normal body mass index and 12% had glucose abnormalities of which 9.5% was for the pre-diabetic type. Equally concerning, during the next year of follow-up, 42% of those with pre-diabetes progressed to probable diabetes, such that by the end of the study 25% of patients with schizophrenia had recognised glucose abnormalities. ^{Reference Kusumi, Ito, Uemura, Honda, Hayashishita and Miyamoto4} Collectively this seems to constitute a strong case for regular cardiometabolic monitoring in high-risk patients, including anyone prescribed long-term antipsychotic medication. Periodic checks were already part of routine care but the literature suggests this practice was inadequate. Systematic monitoring is an improvement but still not adequate on its own. Systematic testing must be tied to clear treatment options and also clear lines of responsibility.