Inhaled Insulin: Dead Horse or Rising Phoenix?

Ten to 15 years ago inhaled insulin (InIn) was a hot topic with high hopes funneling massive investments into this area by small and large pharmaceutical companies. The clinical development of different InIns led to numerous presentations at scientific meetings and many publications, not to mention the hundreds of millions of dollars spent on research and development. This route of insulin administration appeared to have a prosperous future, especially when the first InIn, Exubera, was approved by the FDA/EMA and brought to market in the United States/European Union. A number of other InIn developments were in late stage of clinical development and were expected to quickly arrive in the market. However, Pfizer withdrew Exubera from the market only a few months later due to its limited commercial success, other companies stopped their InIn developments more or less immediately, despite the fact that clinical development showed positive results. Patient preference studies suggested a high acceptance for InIns. Financial losses reached billions of dollars, largely due to Pfizer’s inability to market the bong-like delivery device for this initial InIn.

The only exception and survivor was MannKind, which progressed with their InIn, Afrezza, driven by the resources and will of Al Mann who founded and sold MiniMed. This technosphere insulin (TI) has an ultrarapid onset of action and also a rapid decline in action; it differs in this respect from the other InIns that were tested. The fast pharmacokinetic (PK) and pharmacodynamic (PD) properties of this TI attracted Sanofi and they brought this product to the market in the United States in 2014, but returned the product to MannKind shortly afterward again due to insufficient sales. It appeared that TI’s advantages were insufficient to counterbalance the restricted label, and to a big extent, not getting the word out properly to prescribers and patients both type 1 and type 2 diabetes. Pulmonary function test was another barrier to prescribing TI. It also appears that Sanofi did not promote TI as heavily as needed for a product with the negative history established by its predecessor. Currently, one other InIn is in clinical development by Dance Biopharm that has a more rapid onset of action than current rapid-acting subcutaneously (SC) applied insulin analogs, but has a longer duration of action than Afrezza, improving insulin coverage for mixed meals. Few publications are available for this InIn at this time.

This editorial address the question implied in the title: Can the benefits of InIn counteract its negatives? Can InIn be raised from a near death experience, and if so by and for whom? In other words, the perception that InIns are dead may be premature. ¹

Optimal Coverage of Prandial Insulin Requirements

Whether an InIn is more appropriate for patients with type 1 (T1D) or type 2 diabetes (T2D) is not discussed as this will depend quite a lot on the way given patients handle their insulin regimen. For either type of diabetes, it is important to provide a prandial insulin with a rapid-on and rapid-off metabolic activity is clinically relevant to mimic endogenous insulin secretion in healthy subjects and limit postprandial glycemic excursions. Millions of patients need for better options for reducing postprandial glycemic excursions, HbA1c levels, and glucose variability.

Current prandial insulins, whether delivered via an insulin pump or injected with a pen to the SC tissue, exhibit a relatively slow onset of action and act too long because of the relatively slow absorption from the SC insulin depot. Patients who are aiming for optimal metabolic control and prefer a flexible lifestyle are interested in a prandial insulin coverage that help them to handle meals optimally with respect to timing of insulin with meals, meal size and type, and so on.

The number of attempts to develop products that provide “ultrafast” prandial insulin action has decreased in the last years; for example, no intradermal insulin application made it into a product and also attempts to increase insulin absorption by adding enzyms or other substances (like hyalonoric acid) did not make it.^2,3 So far, only Fiasp by Novo Nordisk has come to the market. ³

Meanwhile, the availability of TI with a proven rapid time-action profile has been ignored. Patients with T1D striving for an optimal metabolic control might adjust doses of an InIn with simultaneous use of a CGM, although respective systematic studies are missing. An InIn is also likely to avoid late postprandial hypoglycemic events with a reduced number of hypoglycemic events. Similarly, use of TI with an artificial pancreas (AP) system may improve postprandial and overall metabolic control. ⁴

Recent research has shown that usage of GLP-1 agonists or SGLT-2 inhibitors improves metabolic control in many patients with T2D. Therefore, if these benefits hold up over time, then the need for usage of prandial insulin in such patients may diminish considerably. However, InIn could be used as adjunct for large meals in patients using GLP-1 agonists, although this requires motivated, flexible patients. Poor adherence and primarily not the lack of good drugs is an issue in many patients. InIn is beneficial in patients with T2D that need an injection of prandial insulin, but are reluctant to use it in practice.

Nevertheless, the need for an ultrarapid prandial insulin in many patients with T1D and at least some subgroups of patients with T2D remains. It will remain to be seen if more widespread usage of AP systems will improve postprandial glycemic control, some approaches use a prandial insulin in multiple daily injection (MDI) users. Usage of an insulin with a more rapid absorption after SC administrations, such as Fiasp, probably offers advantages, but remains to be documented.^3,5 A head-to-head comparison of Fiasp versus TI in—for example—meal studies would be of interest. The absorption of any SC insulin may remain relatively slow and have a duration of action that is too long.

Reasons for Failure of Exubera

In hindsight, the limited commercial success of this InIn had a number of reasons:

• excessive size of the inhaler

• high costs due to low relative bioefficacy (= high costs of goods)

• cumbersome application of sufficient amounts of insulin

• complicated conversion of insulin units into mg insulin to be applied

• inappropriate pre- and postmarketing efforts by the manufacturer, new approaches need time to become successful products

• need for full pulmonary function tests (with TI only spirometry is required, which can be done cheaply and easily)

• the PK and PD properties of the “standard” InIn were comparable to that of rapid-acting insulin analogs applied to the SC tissue. There was no real physiologic advantage in using such InIn (in contrast to TI). Many health care companies required documentation of a needle phobia, which is rare in real life.

A convenient and discrete way to apply insulin (especially in public) is an important feature for patients. Usage of the Exubera inhaler in a restaurant was guaranteed to attract attention. Physician samples were sparse, TV ads were too late and lacked impact, and certified diabetic educators, who play an even bigger role than physicians in putting patients on insulin, were ignored.

Pros and Cons of TI

The time-action profile of TI is characterized by a very rapid onset of action and shorter duration of action than standard InIn and rapid-acting insulin analogs for SC administration. The way decline in action is an advantage if meal with rapidly absorbable carbohydrates are eaten. When eating regular mixed meals, the short duration of action may require a second application of TI or combined therapy with a subcutaneous injection, to avoid an increase in late postprandial hyperglycemic events. On the other hand, this short duration of action makes TI an ideal correction insulin. A patient can apply TI when their glucose is elevated and have less concern with insulin stacking. However, TI’s “units” do not provide an equivalent insulin lowering capability, dose increments are huge for some patients at 4 or 8 “units.” Some patients have a need for smaller increments to optimize their prandial insulin coverage. Patients need to learn the glucose lowering capability of TI’s “units,” which may not reflect SC units in all patients. Four units of TI is like 2.0 to 2.5 U of SC prandial insulin.

For optimal handling of an insulin with a unique time-action profile, in this case with a rapid onset and short duration, an immediate feedback of its impact on glucose control by means of continuous glucose monitoring (CGM) systems is of help, like with SC insulins. However, most patients with T2D don’t do much glucose monitoring in practice. Patients with T1D are much more concerned about optimal glycemic control and avoidance of hypoglycemic events than patients with T2D. Nevertheless, more and more patients on MDI use CGM systems that provide the glycemic insight that is required at a reduced cost and increased convenience (no calibrations). Few clinical studies have been published that show a real advantage for TI, such as improving metabolic control, increase of “time in range,” long-term usage, and so on.

The requirement for pulmonary function testing with spirometry with TI is another burden for patients and clinicians. Although uncommon in endocrinology offices and diabetes clinics, many primary care physicians who see most of the patients with T2D have simple spirometry devices for their asthma and COPD patients.

Reasons for Failure of TI in the Hands of Sanofi

TI offers a number of advantages in comparison to Exubera: a small inhaler, easier and more discrete use, simpler dose calculation, and—most importantly—a rapid onset of action/decline of action.^6,7 So, why did Sanofi fail to succeed?

Like Pfizer, marketing by Sanofi was also sparse for TI. New routes of insulin administration take time to catch on in the market, especially in view of the history of InIn.

A basic misunderstanding was probably that Sanofi regarded TI as the solution for patients who do not want to inject basal insulin; however, such patients are not going to measure their postprandial glycemia several times a day to control a challenging mealtime insulin. According to the label of TI Sanofi could not promote better postprandial glycemic control and less delayed hypoglycemic events 2 to 5 hours after eating.

Also the costs for TI were high and market preparation was not done sufficiently. The difficulty in obtaining reimbursement for such a costly product continues to handicap this market. In view of the increases in insulin prices in the United States it would of be interest to compare the cost per unit of TI with that of the rapid-acting insulin analogs for SC administration. A direct comparison of all costs associated with a given way of delivering prandial insulin via either the SC route or the pulmonary route would be of interest (ie, add costs for pens or pulmonary function test).

In the United States pricing of insulin is complex and involves many parties. ⁸ In part this is also responsible for the double-digit price increases seen with insulin in the past decade. TI is flat priced, making it cheaper than SC prandial insulin if a patient is taking 8-12 U or slightly more if only 4 U of TI are used.

Safety Aspects

A major concern for many diabetologists revolves around the safety aspects of InIns. Insulin’s growth-promoting activity remains a concern for long-term application in the lungs. It may promote cancer in current or former heavy smokers. However, due to the small number of cases there is a high risk of bias, and definitive data are lacking. Other drugs, like the GLP-1 agonists, have a black box warning for people with medullary thyroid cancer or noncancerous multiple endocrine neoplasia 2.

InIn is also known to cause a drop in FEV1 of about 1-2% with insulin powders; however, the mechanism behind this effect (which was seen with most InIn in many, but not all studies) is unclear. It might be, that insulin as an anabolic hormone simply increases the smooth muscle tone and makes the lungs “stiffer.” It is worth to note that the effect was reversible and nonprogressive.

The small potential lung cancer signal seen in the Exubera Clincal studies was only seen in former heavy smokers, and observations were subject to bias. In addition overall survival was better in Exubera patients. However, when Pfizer withdrew their InIn from the market, they published a press statement about lung carcinoma observed in the large clinical studies (that were stopped, which had also positive financial aspects for Pfizer) that exacerbated such concerns and hampers clinical acceptance of another InIn like TI. The high safety requirements for TI in the market approval by the FDA might have been triggered by this. MannKind is doing a huge phase 4 safety clinical trial study – also to study the long term effects of its excipient on the lungs and heart. A potential lung cancer safety signal is too rare to study within the framework of a clinical trial without spending billions; this can be studied by postmarketing surveillance of electronic databases.

Side effects like cough can also hamper patient acceptance. Bronchospasms were observed with TI in 5 out of 17 PK/PD subjects previously diagnosed with asthma, the first time they inhaled it, and InIn is probably not suited for patients with asthma. The incidence of bronchospasm seen with Exubera was much lower.

During the clinical development of different InIns, a number of studies investigated the impact of this route of insulin administration in patients having asthma or a cough or who are smokers. Respiratory factors like these have the potential to increase or decrease absorption of InIn; however, the rapid uptake and elimination of TI was unchanged in smokers.

Studies That Are Missing

With pulmonary administration insulin reaches the blood stream “directly,” that is, the insulin is absorbed rapidly and with an impressively good reproducibility. In contrast, absorption of insulin from the SC depot is known to depend on many factors, such as BMI. It is well known that insulin therapy in obese patients is difficult; due to the fact that insulin is absorbed slowly in such patients, the time-action profile of prandial insulin is shifted toward basal insulin, which means that many of such patients have a bad postprandial metabolic control. However, it was never formally studied whether patients achieved a better postprandial or overall metabolic control if prandial insulin requirements were covered by large doses of InIn rather than by SC insulin administration.

Summary and Outlook

MannKind remains active on the US market with its TI and recently updated its product information for the FDA. It remains to be seen how well MannKind can survive as an independent company with only a single product. The only novel InIn that is in clinical development has some promising properties; however, one has to see if this development leads toward an approved product.

Is InIn a dead horse? I don’t think so. But do I believe that InIn has a bright and straightforward future? To be honest, I’m not sure. I see a role for InIn in treatment of at least certain groups of patients with diabetes. InIn provides good additional options for insulin therapy in patients with diabetes, especially if dosing precision and more convenient selection of a wide range of dose sizes become possible.