We found the incidence of TT1 to be 1/90,102 in Finland, with a significant enrichment (1/9,990) in South Ostrobothnia. The latter is likely due to the homogenous population and overrepresentation of inherited diseases in this area . In central Europe the corresponding figure is ~1/100,000–200,000 [6,19], in Norway 1/74,800  and in Quebec (Canada) up to 1/16,000 , whereas in Japan TT1 is exceptional . As a further supporting founder effect, most patients living in South Ostrobothnia had homozygotic “Finnish type” Trp262X mutation  and their ancestors also originated from there. The other mutations were c.1062+5G>, which is common in Quebec [16,20], c.191delA common in Turkey  and “Mediterranean” c.554-1G>T . Of note, one patient had a previously unreported c.205del (p.Ser69fs) mutation .
Clinical presentations were mostly in line with earlier reports [6,16,17,19,22–27], the main findings including e.g. liver failure, poor growth and rickets. Pregnancies had also been uneventful [17,23], but quite many had temporary hypoglycaemia or hypotonia after birth and five presented with inguinal/scrotum hernias, likely due to ascites . These previously unreported findings should be kept in mind as possible early signs of TT1 in high-prevalence areas. Inexplicably, we found no cases with frequently described [16,19,22,29] cardiomyopathy and neurological crises. Of note, although according to laboratory parameters most of the patients had deep coagulopathy typical for TT1 [22,24,29], severe bleeding was rare.
The number of screen-detected patients was low, but their milder phenotype compared with those detected clinically was evident. Moreover, in line with earlier short-term studies [6,13,22,29], early detection seemed to improve the prognosis; all subjects needing transplantation despite nitisinone had late diagnosis. As regards the benefits of NBS, the current evidence is again based mainly on short-term reports [6,17,29], an exception being a Canadian study which found none of the NBS patients to have significant liver problems after 5–10 years . However, like some other groups [6,17,29]we observed signs of hepatic dysfunction which, together with the aforesaid neonatal symptoms, suggests disease progression already in utero [17,30]. The frequency of neonatal hypoglycaemia was also surprisingly high. Although this could be due in part to the sensitive screening performed frequently on Finnish newborns, it may also be TT1-related, and further studies on this interesting issue are needed. Risk of later health problems emphasizes the need for careful follow-up also for screen-detected patients .
Basic laboratory values and liver function tests normalized rapidly in most cases, while this took longer in case of transaminases and biliary parameters. This is mostly in line with earlier reports [17,19,23,25], although the follow-up time has usually been shorter. Transaminase levels were higher and normalized more slowly than described before [16,23], but this did not predict later complications. Some children also presented with high ammonium ion (NH4+), this being feared to predict early need for transplantation , but the values decreased promptly on nitisinone. Slow decline of AFP is a physiological phenomenon in infancy [30,31], but our findings confirm that nondecreasing or rerising values predict malignancy [16,19,22,32]. Of note, persistent thrombocytopenia, which was associated with splenomegaly and is a classical sign of a chronic liver disease, was also a strong predictor of subsequent need for transplantation.
The kidney, CNS, cardiac and bone imaging findings normalized within a few years, but some patients had persistent liver abnormalities and/or splenomegaly. Comparable gradual improvement in renal findings has also been reported by a British group  whereas clinically important data about the disappearance of the other findings or significance of their perseverance has been limited. Persistent liver abnormalities have been reported in 19–57% of patients, but the follow-up times have been shorter than in the present study [6,16,17,19,23,29]. Here the nonresponsive findings and reappearance of splenomegaly were major warning signs for subsequent hepatic transplantation.
Long-term complications of TT1 were mostly analogous with those reported in the literature [15–17,19,25,29,33,34]. Neurological problems and poor growth were more common here, but the follow-up times of these earlier studies may have been too short to detect these late-appearing issues. The former have been suggested to be partially attributable to the side effects of nitisinone [34,35], but this is debatable [33,36]and here low levels seemed more harmful. However, early disease onset seems to increase the risk for neurological complications , again supporting the idea of in utero progression of TT1. Poor growth has been sparsely reported28, but the possible risk for short stature observed here calls for further studies. The association observed between low mean nitisinone and growth failure may be due to inadequate treatment and ongoing liver disease. Then again, these same patients needed liver transplant despite nitisinone use, and hence the role of transplantation must also be considered. There is also limited and inconsistent data about the prevalence and appropriate follow-up of bone issues in TT1 [16,12], but our results suggest that surveillance at least for cases with rickets at diagnosis is warranted. In contrast, as in prior short-term studies [17,19,23,25,29], persistent renal involvement seems to be rare.
We found no significant association between long-term complications and tyrosine levels, but low serum nitisinone, even with negative urine SA, was associated with learning difficulties, growth delay and a need for transplantation. The reported nitisinone target range has varied 20–80 µmol/l [6,11,37]and the dosing usually aims to achieve negative urine SA [6,24]. Optimal levels remain somewhat unclear, but frequent monitoring of the levels with a target of 40–60 µmol/l has been recommended . Interestingly, recent studies have reported increased blood SA with nitisinone concentrations <44.3 µmol/l  and positive urine SA with nitisinone <40 µmol/l, respectively . Thus, there may in theory have been intermittent SA secretion in our patients with low mean nitisinone levels even without detectable urine SA at follow-up visits. Urine SA varies depending on urine concentration and blood measurement may be more stable and replicable indicator of ongoing SA production . Thus, the latter in combination with the nitisinone level could thus be preferable as a follow-up marker [38,39,40]. Interpretation of the markedly fluctuating nitisinone values can be difficult, but our results underline the importance of sufficient levels, which may be even higher than previously suggested. Of note, the increased ratio of natural to modified protein was also associated with subsequent need for liver transplantation in nitisinone-treated patients. This may be related to generally higher tyrosine levels, although we found no statistically significant association between tyrosine levels and complications.
The main strengths of the present study were the nationwide coverage and availability of comprehensive medical data. Detailed registers also provided information on long-term outcomes and enabled us to assess risk factors for later complications, although weaknesses in the retrospective design remain. As a further limitation, an intensified register search was conducted in only one district, which in theory could have led to a few earlier cases being missed. Some patients may also have died before the transplantation era and were thus lost since medical records are deleted 12 years posthumously.