medwireNews: Further analysis of the French cohort of the SAGhE study raises the possibility of a link between growth hormone (GH) treatment in childhood and an increased risk of haemorrhagic stroke in early adulthood.

The absolute number of strokes was small, however, with 11 occurring among a cohort of over 6000 patients, compared with an expected number of 3.1 to 7.4, calculated from population-based stroke registries.

Commenting on the study for medwireNews, Martin Savage, Emeritus Professor of Paediatric Endocrinology from Barts and the London School of Medicine & Dentistry, UK, stressed that a causative role of GH for haemorrhagic stroke “remains highly speculative and as yet unproven.”

The researchers suggest that GH treatment may lead to weakened blood vessel walls, as seen in acromegaly. But Savage argues that the low doses of GH used “would be unlikely to cause pathologically elevated insulin-like growth factor-1”, which, combined with the short treatment duration, makes acromegaly an “inappropriate” comparison.

The multinational SAGhE (Safety and Appropriateness of Growth hormone treatments in Europe) study monitors people who received GH during childhood in the late 1980s and 1990s and involves patients from eight European countries. The French team previously reported an increased mortality risk among patients in the French cohort; however, a separate analysis of the Belgian, Dutch and Swedish cohorts found no such increase, reporting a low number of deaths, which were predominantly accidental.

The French investigators – Joël Coste (Université de Lorraine and University Paris Descartes, Paris) and team – found the largest excess mortality risk in their cohort to be for subarachnoid or intracerebral haemorrhage, and have now investigated this further in a report published in Neurology.

However, Savage said that the public concern in the wake of the initial French report “should have persuaded Coste and co-authors to report their results together with the findings of the full Consortium, rather than publish individually for the second time.”
He added: “One has to question the motives for this ‘go-it-alone’ approach which will ultimately be seen to damage rather than enhance the value of this important collaborative project.”

The analysis of the Belgian, Dutch and Swedish cohorts is not mentioned in the second French report, which Savage described as “an extraordinary omission, given the collaborative nature of the Consortium study.” He said: “This adds to the unbalanced nature of this publication.”

As detailed in the report, there were 11 cases of stroke among 6874 patients in the French cohort defined as “low risk”, after exclusion of those treated for severe conditions such as cancer, and those with multiple pituitary hormone deficiency or paediatric syndromes such as Turner, Prader-Willi or Fanconi. The strokes comprised three ischaemic strokes, three intracerebral haemorrhages and five subarachnoid haemorrhages; six were identified in patients’ medical records and five from patients’ responses to questionnaires.

Stroke occurred at an average age of 24 years among the patients. Three had been born small for gestational age, but none had severe GH deficiency.

The expected number of strokes was 7.4, based on a population-based stroke registry from Dijon, France, giving a nonsignificantly increased crude standardised incidence rate (SIR) of 1.5 and a significantly increased corrected SIR (including estimated number of missed cases) of 2.2.

The crude SIR was significantly increased for subarachnoid haemorrhage (5.7) and nonsignificantly increased for intracerebral haemorrhage (2.1), and the corresponding corrected SIRs were significantly increased for both, at 8.6 and 3.9. All SIRs were slightly higher when a stroke registry from Oxford, UK, was used as the comparator.

By Eleanor McDermid, Senior medwireNews Reporter

Neurology 2014; Advance online publication

Free Abstract