Our study suggests an association between TS and MTHFR polymorphisms and the appearance of d4T-related toxicity in the form of acute pancreatitis or peripheral neuropathy. However, our work has inherent limitations. First, this is a case-control study and therefore no causal relationships should or must be drawn. Second, in case-control studies endpoint verification is of paramount importance. It is known that peripheral neuropathy or acute pancreatitis in the setting of HIV-1 infection treated with HAART may be caused not only by antiretroviral drugs but also by other drugs, alcohol abuse, or biliary stones and hypertriglyceridemia in the case of pancreatitis 
. On the other hand, there is no pathognomic test to etiologically ascribe the etiology of peripheral neuropathy or pancreatitis. Notwithstanding that, none of these well-known causes was present when patients in the study were diagnosed as having peripheral neuropathy or pancreatitis, except for alcohol abuse which in fact was more frequent among controls. Unfortunately, when the study was performed, intracellular d4T-TP concentrations in PBMCs were not available. Third, d4T is a well-known cause of peripheral neuropathy and pancreatitis, especially when combined with didanosine (ddI) or even more when combined with ddI and hydroxyurea 
. Among our cases, 23 (53.5%) were also exposed to ddI, which may have contributed to the development of pancreatitis and/or peripheral neuropathy. Although cumulated exposure to ddI was not different in cases and controls, taking d4T plus ddI at event was statistically more frequent in cases. However, taking both drugs in combination was not an independent predictor for the development of peripheral neuropathy and/or acute pancreatitis. Fourth, there are a number of gene markers which have been associated with the risk of developing both toxic peripheral neuropathy and pancreatitis in HIV-infected patients on HAART, such as cystic fibrosis trans-membrane conductance regulator (CTR) and serine protease inhibitor kazal-1 (SPINK-1) mutations for pancreatitis and mitochondrial haplotype T for peripheral neuropathy 
. We cannot exclude the possibility that some of these untested genes may have contributed to acute pancreatitis and/or peripheral neuropathy in our patients.
Acute pancreatitis in the setting of HIV infection and antiretroviral therapy has wide incidence rates ranging from 1.27 to 22.6events/1000 PY 
. These wide incidence rates may be explained by different diagnostic criteria (clinical vs. laboratory-based), and by the fact that some of the studies were performed in early calendar years, when more toxic drugs were used. Its appearance has been linked to the classical risk factors; i.e. alcohol abuse, hypertriglyceridemia, as well as to the use of antiretroviral drugs such as d4T and ddI 
. The incidence rates for peripheral neuropathy have also been very wide, ranging from 0.7 to 39.7/1000 PY 
. Known risk factors for developing peripheral neuropathy in HAART-treated patients include alcohol abuse, treatment with other neuropathic drugs (isoniazid, methotrexate,…), factors which were ruled out in our patients 
. However, sometimes it is difficult to distinguish toxic peripheral neuropathy from exacerbation of HIV-associated neuropathy in HAART-treated patients, which may also contribute to these varying incidence rates. Notwithstanding that, none of our patients had clinical features consistent with peripheral neuropathy prior to taking d4T. The incidence of both acute pancreatitis and peripheral neuropathy in our work fall within these wide ranges, most probably because our study period spans throughout a long time period when mitochondrially-toxic drugs were still widely used.
Acute pancreatitis has been associated with mutations in CTR and SPINK-1 genes both in the general population and in HIV-infected patients 
. Among HIV-infected patients with acute clinical pancreatitis in a small study, 40% were carriers of CTR or SPINK-1 mutations 
. Similarly, a number of gene polymorphisms have been described, associated with an increased risk of developing peripheral neuropathy associated with d4T and ddI use, including mitochondrial haplotype T 
, the mitochondrial polymorphism MTND2/HON4917G, specific to haplogroup G 
, and the 282 C>Y hemochromatosis gene mutation 
Among the factors independently associated with the development of neuropathy or pancreatitis, we found a low CD4 cell count nadir, a common marker for most toxicities and co-morbidities 
. Most probably it is a signal of increased inflammation and/or immune activation, which in turn contribute to overexpression of nucleoside transporters potentially leading to increased intracellular d4T-TP levels 
. The association of TS and MTHFR polymorphisms with mitochondrial toxicity has biological plausibility, since genotypes which have a negative impact on TS activity are associated with higher d4T-TP intracellular levels and with fat redistribution, a well-known toxicity of mitochondrial origin 
. Not surprisingly, we found that low expression TS genotypes are also associated with the appearance of d4T-induced mitochondrial toxicity in the form of peripheral neuropathy whereas the association for acute pancreatitis showed only a trend to statistical significance, most likely because of the small number of cases. A potentially contributing factor could have been that cases had a higher per weight exposure to d4T than controls, but again this was not an independent predictor of development of neuropathy or acute pancreatitis.
MTHFR gene is also polymorphic. The best known polymorphism consists of a 677 C→
T transition in exon 4, which results in an alanine to valine substitution in the predicted catalytic domain of MTHFR. This substitution renders the enzyme thermolabile and homozygotes and heterozygotes have about 70% and 35% reduced enzyme activity, respectively 
. A second common polymorphism in the MTHFR gene is a 1298 A→
C transition in exon 7 which results in a glutamate to alanine substitution within a presumed regulatory domain of the protein 
. The 1298C allele leads to a decreased enzyme activity, and individuals who are compound heterozygous for the 677T and 1298C have a 40–50% reduced MTHFR activity 
. Therefore, the non-mutated forms of MTHFR exhibit high enzymatic activity and should lead to high levels of intracellular 5,10-methylene tetrahydrofolate that will favor the formation and stability of the inhibitory ternary complex involving TS, 5,10-methylene tetrahydrofolate and deoxyuridine monophosphate 
. Consequently, patients with non-mutated alleles should have a decreased TS activity, which in turn may lead to increased d4T-TP intracellular levels and to clinical toxicity including pancreatitis and peripheral neuropathy.
Stavudine-associated side effects may seem of only limited relevance today, because its use has greatly decreased in developed countries mainly due to its association with fat distribution abnormalities 
. However, its use in developing countries as part of the starting antiretroviral regimes, with its associated toll of toxicities 
, is still common despite the WHO recommendation to substitute for less toxic NRTI when feasible 
In summary, our study suggests that d4T-associated acute pancreatitis and/or peripheral neuropathy are associated with low-degree TS expression and MTHFR genotype associated with an increased enzymatic activity. There is a plausible pathogenic mechanism for such an association since it is well-known that, in d4T-treated patients, the presence of low-degree TS expression genotype is associated with increased d4T-TP intracellular concentrations and MTHFR genotype associated with an increased enzymatic activity contributes to a decreased TS functionality. This may be of value in tailoring d4T therapy when needed.