There is a poor correlation between measurable immune responses to hookworms, the intensity of gut inflammation, and the longevity of hookworms which, in the case of A. caninum in the human intestine, is consistently and markedly limited.Parasite survival depends on a multitude of factors, beginning with an ability to negotiate host anatomy. That larvae of canine hookworms can read internal human signposts is remarkable, but does indicate molecular similarities between tissues of different hosts. When L4 arrive in the intestine, they anchor and feed normally initially, but then seem to run out of energy, to succumb at about three weeks. Immune and inflammatory responses at about this time are not particularly unique or aggressive, for the event is clinically silent in most infected people.Croese accepts that `other hookworms are smaller than normal when found outside the specific host', and that `rapid [hookworm intestinal] transit would restrict the possibilities for the digestion and absorption of complex constituents'[1xSee all References[1]. He acknowledges that the worm might not be able to derive sufficient nutrient from its host. By assuming that parasite nutrition depends on simple molecules absorbed from host blood, he forgets that the worm's digestive enzymes commence working externally, to predigest mucosal tissues. Further, while transit through the worm intestine seems rapid from our perspective, we must remember the scale at which all this is happening! And the worms do not pump food continuously; periods of rest allow for more complete digestion and absorption of material within its own gut.This begs the question: how compatible are parasite enzymes with their substrates, ie. host tissues? If hookworm digestive physiology is working at the limits of its capacity to support the rapidly growing juvenile, the enzymes that predigest host tissues, and continue the process in the worm's gut, must function optimally. A modest reduction in host-mucosal breakdown or haemolysis (assuming blood and mucosa are equally important nutrient sources) could jeopardize parasite development and survival (perhaps via resistance to inflammatory damage?). If constituent proteins in canine gut mucosa and blood vary slightly from those of humans, the corresponding hookworm enzymes would be expected also to vary slightly. Such differences might be indiscernible by immunological probes, but possibly not by molecular and enzyme kinetic studies. Notwithstanding the cross-species activity of secreted hookworm anticoagulants and NIF, Croese acknowledges that in these closely related parasites, `It seems probable that some excretory/secretory molecules are subtly and crucially unique, and that these differences determine host specificity'.In effect, the worm starves. This might also lead to a form of `immunosuppression' in the worm, ie. a decline in output of host-immunomodulating factors (and so indirectly allow intense EE in some patients).Congruity between host proteins and parasite enzymes might underlie host-specificity in a wide range of haematophagous parasites, but has not been examined in this context.It remains to explain the experimental establishment of hookworms in abnormal hosts. While A. ceylanicum, normally parasitic in dogs and cats, has been established in hamsters, only neonatal animals are susceptible. Apart from immature immune systems, neonatal mammals have high levels of foetal haemoglobin, which might vary less between species than does adult haemoglobin. A. duodenale has been established in dogs, which initially were immunosuppressed. While this might suggest that immunity is important, immunosuppressive therapy, especially that involving corticosteroids, has a diversity of biochemical and physiological effects on most body systems, which indirectly may favour hookworm survival. Steroids might even mimic helminthic ecdysterones, and so directly stimulate growth and development of some parasites, as has been speculated for Strongyloides stercoralis[8xGenta, R.M. Clin. Microbiol. Rev. 1992; 5: 345–355PubMedSee all References[8].Finally, it must be recognized that successfully established laboratory models of A. duodenale in dogs represent selected strains of both parasite and host. This `directed evolution' does not need immunological explanations. If the strains of A. caninum that develop in the human intestine could be subcultured in immunosuppressed children, before too long we should have a successful model of canine hookworm infection in human hosts!