Mecasermin (Increlex)

Identity and regulatory status

Mecasermin is recombinant human insulin-like growth factor-I: seventy amino acids in a single chain, three intramolecular disulfide bridges, produced in Escherichia coli carrying the human gene. The FDA Global Substance Registration System holds it under UNII 7GR9I2683O with CAS 68562-41-4 as the primary registry code and 67763-96-6 recorded on the same substance as a generic family code, alongside INN 6372, USAN HH-78 and ATC code H01AC03. The disulfide pattern is not in that record, which carries no disulfide links at all; it is annotated on the mature IGF-I chain in UniProt P05019. Rinderknecht and Humbel determined the sequence in 1978 and calculated a molecular weight of 7649; that is the figure both the FDA and the EMA print today. PubChem holds no compound record, and queries on the name and on both registry numbers return substance deposits only.

The FDA approved Increlex on 30 August 2005 under BLA 021839, for growth failure in paediatric patients aged two years and older with severe primary IGF-1 deficiency, or with growth hormone gene deletion who have developed neutralising antibodies to growth hormone. The EMA authorised it on 2 August 2007 under exceptional circumstances, the designation applied when an applicant cannot supply comprehensive data on efficacy and safety under normal conditions of use. Its EU orphan designation was removed from the register in August 2017 at the end of the ten-year exclusivity period. No adult indication exists in either jurisdiction.

A second product, mecasermin rinfabate, an equimolar complex of rhIGF-I with its binding protein IGFBP-3, was approved by the FDA on 12 December 2005 under BLA 021884 and appears in Drugs@FDA as discontinued. Its withdrawal followed a March 2007 settlement of patent litigation brought by Tercica and Genentech, described in the sponsor's own filings with the Securities and Exchange Commission. The two products are routinely conflated in secondary summaries, which matters because the binding protein is what governs clearance, so their pharmacokinetics are not interchangeable.

The trials behind the approval

One randomised, double-blind, placebo-controlled trial exists for the approved indication. Guevara-Aguirre and colleagues enrolled seventeen prepubertal children with growth hormone receptor deficiency from a genetically homogeneous population in southern Ecuador. Subjects received either twelve months of subcutaneous rhIGF-I or six months of placebo followed by six months of active treatment. Growth rate in treated subjects rose from 2.9 plus or minus 0.6 to 8.6 plus or minus 0.4 cm per year. Hypoglycaemia occurred at equal frequency in the placebo and rhIGF-I groups, and one recipient developed papilloedema that resolved without intervention.

Chernausek and colleagues reported seventy-six children with IGF-I deficiency due to growth hormone insensitivity, treated for up to twelve years under a predominantly open-label design, which is the design of every study in the programme larger than the Ecuadorian trial. Height velocity rose from a mean 2.8 cm per year at baseline to 8.0 cm per year in the first year, varied with the dose administered, and remained above baseline for up to eight years while declining each year. The European label tabulates the pooled programme, ninety-two subjects across five studies, eighty-one evaluable, and records mean height velocity falling from 8.0 cm per year in year one to 4.4 cm per year in year eight.

Backeljauw and colleagues followed twenty-one treatment-naive patients to adult or near-adult height over a mean of ten years. Cumulative mean change in height standard deviation score was plus two, and the observed height gain was 13.4 cm more than had been expected without treatment. Nine of the twenty-one also received a gonadotropin-releasing hormone analogue for a mean of 2.9 plus or minus 1.8 years, and the paper records the adult height of that subgroup as no different from those given rhIGF-I alone. The comparator is an expectation derived from untreated Laron syndrome rather than a randomised control arm, and the authors record that most patients did not reach the normal adult range. The European label states the same result as approximately 13 cm plus or minus 8 cm after an average of eleven years, so the label and the paper it derives from disagree on both the figure and the duration.

What the registries record

Bang and colleagues reported the European Increlex Growth Forum Database, an observational registry of clinical-practice use covering 195 enrolled children between December 2008 and September 2013. Mean height velocity in the first treatment year was 6.9 plus or minus 2.2 cm per year across the 144 patients with data, and 7.3 plus or minus 2.0 cm per year in the eighty-one prepubertal patients naive to treatment. Both sit below the 8.0 cm per year recorded in year one of the pooled trial programme, on a population selected by clinical practice rather than by protocol.

Ramon-Krauel and colleagues reported the global version of the same registry in 2026, covering the 102 patients who had reached near-adult height by the April 2023 data cut. Mean age at initiation was 11.8 years and median treatment duration 3.9 years. Mean height standard deviation score gain was 0.9 plus or minus 1.1 overall and 1.4 plus or minus 1.0 in treatment-naive prepubertal patients. Among treatment-naive prepubertal patients, 46.7 per cent, twenty-one of forty-five, reached a near-adult height inside the normal range; among patients with Laron syndrome, 10.5 per cent did, two of nineteen.

Treatment naivety predicted height gain in the overall near-adult-height population, and the investigators record the greatest improvements in the treatment-naive prepubertal group. Late initiation is the dominant feature of the registry data: a mean starting age of 11.8 years leaves little growth remaining, and the investigators attribute the outcomes in the wider cohort to it. Neither registry carries a control arm, and neither was designed to separate the contribution of the drug from the natural history of the underlying condition.

Adverse events, and the neoplasia contraindication

Hypoglycaemia is the event recorded most often. Chernausek's cohort reported it in 49 per cent of treated subjects; the FDA label, drawing on seventy-one subjects and 274 subject-years, records 42 per cent, with severe episodes in five subjects and hypoglycaemic seizures or loss of consciousness in four. Of the thirty subjects reporting hypoglycaemia in the label dataset, fourteen had a documented history of it before treatment began, and the one randomised trial recorded equal frequency on placebo and on drug. Attributing the whole of that signal to the compound does not survive the comparison.

Lymphoid tissue hypertrophy comes next. Tonsillar or adenoidal hypertrophy appeared in 22 per cent of Chernausek's subjects and 15 per cent of the label cohort, mostly within the first one to two years; seven subjects in the label dataset underwent tonsillectomy or adenotonsillectomy, three of them with obstructive sleep apnoea that resolved after the procedure. Injection-site lipohypertrophy ran at 32 per cent in Chernausek's series. Intracranial hypertension occurred in three subjects. Registry rates are lower across the board, hypoglycaemia 17.6 per cent, lipohypertrophy 10.6 per cent, tonsillar hypertrophy 7.4 per cent, which is what the difference between protocol monitoring and clinical-practice reporting would predict.

In December 2019 the sponsor, in agreement with the EMA and the MHRA, issued a Direct Healthcare Professional Communication stating that a higher number of benign and malignant neoplasms had been identified in patients receiving mecasermin after marketing than the background incidence in that population, that the cases spanned a variety of malignancies including rare ones not usually seen in children, and that most had arisen outside the authorised indication or above the authorised maximum. The European contraindication was widened to any child or adolescent with active or suspected neoplasia, or any condition or history raising the risk of one. The communication gives no numerator, no denominator and no rate.

Half-life is a property of the recipient

The figure quoted almost everywhere is 5.8 hours. Both the FDA label and the European summary of product characteristics attribute it to the mean terminal half-life of total IGF-1 after a single subcutaneous dose in three paediatric subjects with severe primary IGF-1 deficiency, aged twelve to twenty-two. The half-life column of that table carries an n of three. The same table reports a coefficient of variation of 64 per cent on that half-life, and draws apparent clearance and volume of distribution from twelve subjects rather than three.

Grahnen and colleagues measured the same peptide in two populations in 1993 and obtained two answers. In healthy volunteers, clearance was about 0.20 mL per minute per kilogram and half-life about twenty hours. In patients with growth hormone receptor deficiency, clearance was about 0.60 mL per minute per kilogram and half-life about six hours. Subcutaneous bioavailability was 100 per cent in the healthy volunteers; the report gives no bioavailability figure for the growth hormone receptor deficiency arm, and the FDA label records that absolute bioavailability in severe primary IGF-1 deficiency has not been determined. The mechanism is stated in the label itself: more than eighty per cent of circulating IGF-1 travels as a ternary complex with IGFBP-3 and an acid-labile subunit, IGFBP-3 is low in severe primary IGF-1 deficiency, and clearance is inversely proportional to it.

A third set of figures, ten to twelve minutes free, twenty to thirty minutes in the 50 kDa complex, twelve to fifteen hours in the 200 kDa complex, comes from Guler and colleagues, who injected radiolabelled tracer into two normal adults in 1989. Those are tracer kinetics of the endogenous peptide across carrier fractions rather than the disposition of an injected therapeutic dose. Four defensible half-lives exist for one molecule, and each belongs to the population it was measured in, so a page printing one of them without its population is describing something narrower than it appears to.

The indications that did not survive a trial

Amyotrophic lateral sclerosis absorbed the largest effort. Lai and colleagues randomised 266 patients across eight North American centres to placebo or one of two subcutaneous doses for nine months, and reported progression of functional impairment on the Appel scale 26 per cent slower at the higher dose than on placebo. Borasio and colleagues randomised 183 patients across eight European centres over the same nine months and found no significant difference on the same primary endpoint. Sorenson and colleagues then randomised 330 patients at twenty centres for two years and found no difference in manual muscle testing, tracheostomy-free survival or the revised ALS functional rating scale.

Pooling the two nine-month trials, the 2012 Cochrane review reported a mean difference of minus 4.75 on the Appel scale, 95 per cent confidence interval minus 8.41 to minus 1.09, favouring treatment, while recording that every secondary outcome showed only non-significant trends. The same review rated all three included trials at high risk of bias and graded the evidence underlying that pooled estimate as low quality. The development name used for that programme, Myotrophin, survives in the FDA substance record as a synonym on the same entry as Increlex. No regulator approved the molecule for the indication.

Rett syndrome followed a similar arc. Khwaja and colleagues ran a phase 1 study in twelve girls with MECP2 mutations, nine of them with Rett syndrome, recorded no hypoglycaemia or serious adverse events, and described improvement in apnoea and in anxiety and mood measures during a twenty-week open-label extension with no placebo group. O'Leary and colleagues then ran a placebo-controlled crossover trial in thirty girls with classic Rett syndrome and found no significant difference between randomisation groups on any primary endpoint; the Kerr severity scale, the ADAMS depressed mood subscale and several EEG parameters moved toward worsening. That paper carries a corrigendum adding an omitted author and two omitted grant acknowledgements, and no result changed.

The human muscle data

Human data on rhIGF-I and skeletal muscle exist, and they are narrower than the summaries built on them. Fryburg and colleagues studied twenty-two postabsorptive adults during hyperaminoacidaemia, of whom twelve received IGF-I into one brachial artery and ten received insulin, and measured phenylalanine balance and kinetics in both forearms. In the twelve given IGF-I, the infused arm showed a greater positive shift in phenylalanine balance than the contralateral control arm, and the mechanism separated from insulin's: IGF-I acted by stimulating protein synthesis, insulin by inhibiting degradation. That is a six-hour local infusion study of protein turnover, not a study of muscle mass.

Thompson and colleagues randomised sixteen healthy elderly women to growth hormone or one of two rhIGF-I doses for four weeks on a standardised diet. There was no placebo arm. Lean body mass and nitrogen retention rose significantly in the growth hormone group and the higher rhIGF-I group, and fat mass fell in all three groups. The same two groups recorded headaches, lethargy, joint swelling and pain, and bloating. Five or six women per arm over four weeks is the size of that dataset, and it remains the closest published approach to a body-composition study of this peptide in adults without a growth disorder.

Rutter and colleagues randomised thirty-eight prepubescent ambulatory boys with Duchenne muscular dystrophy already receiving glucocorticoids, seventeen to subcutaneous rhIGF-1 and twenty-one to glucocorticoid alone, for six months. Change in six-minute walk distance, the primary endpoint, was 3.4 plus or minus 32.4 metres in the treated group against minus 5.1 plus or minus 50.2 metres in controls, p equal to 0.53. Height velocity was 6.5 plus or minus 1.7 against 3.3 plus or minus 1.3 cm per year, p below 0.0001, and lean mass and insulin sensitivity rose in the treated boys. The authors conclude that six months of treatment improved linear growth and did not change motor function (PMID 32108355).

What is not known

The approved indication rests on one randomised placebo-controlled trial of seventeen children whose placebo arm ran six months. Everything else supporting it is open-label or observational, and the pooled registration dataset is ninety-two subjects. No trial has compared mecasermin against growth hormone in a population able to respond to both, and the near-adult-height figure most often quoted is measured against a projection rather than a control arm, in a cohort of twenty-one of whom nine also received a gonadotropin-releasing hormone analogue. The neoplasia signal that widened the European contraindication in 2019 has no published rate, no denominator and no follow-up duration attached to it, and the EMA authorisation remains one granted under exceptional circumstances, meaning the agency itself recorded that comprehensive efficacy and safety data could not be supplied. Pharmacokinetics are characterised in three subjects for the half-life the labels print, and the labels state that the pharmacokinetics have not been studied below age twelve or above age sixty-five, and that no studies have been conducted in renal or hepatic impairment. On pregnancy and lactation the labels record no human data at all: the FDA label carries animal reproduction data only, a rabbit no-observed-adverse-effect level of 0.5 mg per kilogram per day for fetal toxicity with an increase in fetal death at 2 mg per kilogram, and states that exposure during pregnancy is unlikely because the drug is not indicated after epiphyseal closure. The EMA text describes the animal studies as insufficient with respect to reproductive toxicity, instructs that the product should not be used during pregnancy, recommends a negative pregnancy test and adequate contraception in women of childbearing potential, and does not recommend breast-feeding because information on excretion in human milk is insufficient. For every use outside the growth indication the record is either negative or absent: of three randomised trials in amyotrophic lateral sclerosis two found no difference on their primary endpoint and the third was rated at high risk of bias along with the other two, and one crossover trial in Rett syndrome and one controlled trial in Duchenne muscular dystrophy failed their primary endpoints. No placebo-controlled randomised trial of the compound in healthy adults for muscle mass, athletic performance, injury recovery or ageing was located in ClinicalTrials.gov or PubMed; the only randomised body-composition data are Thompson 1995, sixteen healthy elderly women across three active arms over four weeks, with no placebo group. Material sold outside the approved supply chain has not been subject to the identity, purity or sterility controls that apply to a licensed biologic, and no published analysis of such material was located.

Questions