Intended for healthcare professionals


Inhaled insulin

BMJ 2004; 328 doi: (Published 20 May 2004) Cite this as: BMJ 2004;328:1215
  1. Stephanie A Amiel, professor
  1. Department of Medicine, King's College School of Medicine, London SE5 9PJ
  2. Department of Endocrinology and Metabolic Medicine, Imperial College, St Mary's Hospital, London W2 1NY
  1. K George, professor,
  2. M M Alberti (george.alberti{at}
  1. Department of Medicine, King's College School of Medicine, London SE5 9PJ
  2. Department of Endocrinology and Metabolic Medicine, Imperial College, St Mary's Hospital, London W2 1NY

    May prove to be a panacea

    For over 80 years exogenous insulin has been given by injection. The injection devices have improved—disposable syringes and pen injection devices are more convenient and less traumatic than the boil to sterilise, use until too blunt devices of yesteryear—but patients and healthcare professionals remain uneasy about the concept of injections. Yet the evidence based drive for increasingly tight glycaemic control means that more patients should be offered more injections. A recent attempt to circumvent the need for injection that may soon hit a clinic near you is the use of the lung as an absorption pathway, with the development of insulins to be taken by inhalation. Two versions, a powder and an aerosol, may be nearing launch.

    Insulin can be effective given by inhalation. This was first shown in 1971, although the early work was not pursued, and it was not until 2000 that the modern era of inhaled insulin began.1 2 The bioavailability is 10-15% and the dose equivalent about three times that of injected insulin. The pharmacodynamics of inhaled insulin offer an action profile with a fast onset (although slightly longer run-off) akin to that of rapid acting insulin analogues given subcutaneously, which in studies have shown better postprandial glucose control and less tendency to nocturnal hypoglycaemia.2 3 A Cochrane review of randomised controlled trials comparing inhaled with injected short acting unmodified human insulin used for prandial insulin replacement, in conjunction with a basal injected insulin, concluded that inhaled insulin provided equivalent control to fully injected regimens.4 However, the injected regimens were not always optimised: analogue insulins were not included, and control groups often simply continued their pre-trial regimens of injection twice daily. An opportunity to improve control with injected insulin was missed. With either regimen, fewer than 25% patients achieved optimal control. The included trials were short and conducted in patients with type 1 or 2 diabetes already on insulin. The review did not examine trials of inhaled insulin added to oral agents, but the data would again imply bioequivalence. In patients with type 2 diabetes, adding inhaled insulin to oral hypoglycaemic regimens does improve control more than doing nothing.5

    The advantages of inhaled over injected insulins to date relate to patients' preferences. This is important—apart from patients' comfort, an expensive new insulin could have huge potential advantage if it encouraged adherence and resulted in more patients with diabetes achieving treatment targets. Sadly, the published data on patients' satisfaction, superficially encouraging, are difficult to interpret, as invariably patients have been comparing a new treatment with an old one. When a specially designed questionnaire was used, treatment satisfaction improved significantly in patients with type 1 and type 2 diabetes on taking part in the trials, irrespective of whether the mealtime insulins they took were injected or inhaled.68 Although the improvement was greater with inhaled insulin the injected insulin treatment was identical to pre-trial treatment limiting the potential for improvement. Notably, improvement in treatment satisfaction correlated with improved glycaemic control. Might greater satisfaction have been obtained with injected regimens if these had been optimised effectively? Were the studies just too short to show the biomedical gains one might anticipate from a treatment expressly designed to support compliance, or do the problems of insulin therapy extend beyond a dislike of needles?

    Inhaled insulin has potential problems. The bioavailability is affected by asthma (decreased) and smoking (increased).9 10 Of course, if patients really dislike injections so much inhaled insulin might make its biggest impact on complications of diabetes if it were to be available only to proved non-smokers. Formation of anti-insulin antibodies is higher with inhaled insulin, and although this is dismissed as not affecting insulin requirement over time, older diabetologists will remember the drive to reduce insulin antibody formation, with the fears that antibodies delay and render unpredictable insulin absorption and even that antibody-antigen complexes may increase risk of microvascular disease.11 Finally, there have been concerns about possible long term effects of insulin on lung structure and function, although current published trials report no deleterious effects over the short-term.

    Few people like injections and some are so terrified they refuse appropriate treatment for diabetes. Many are discouraged by previous experience of injections—none of whom will have used current insulin injection devices—and by healthcare professionals using injection therapy as a threat in a (vain) attempt to improve adherence. For patients with established type 1 diabetes lack of freedom to eat or not eat and the demand for (painful) blood glucose testing may be much more of an issue than injection therapy itself.12 In one study only 14% of injections were missed because they were injections.13 In this group healthcare providers will need more robust evidence of patients' preference than is currently available. Where inhaled insulins could really have an impact (in the developed world) will be if healthcare professionals and patients start to use insulin much earlier and more aggressively in type 2 diabetes, affecting the progression of diabetic complications. In the developing world, where cultural taboos against injection treatments may be even more real than here, inhaled insulin may be expected to deliver more in terms of health benefit—but is not likely to be more affordable or available than the currently inadequate supplies of injected insulin. Meanwhile, all patients are waiting to see if the new inhalations are safe. If they are, and if they are cheap enough, at least one barrier to better diabetes treatment may fall.


    • Competing interests KGMMA has received consultancy fees from Pfizer and non-specified research funds from NovoNordisk; SA has received consultancy fees from Aventis, Pfizer, and NovoNordisk for use in research.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    View Abstract

    Log in

    Log in through your institution


    * For online subscription