The aim of this literature review is to summarize and discuss the available evidence for a relationship between polymorphisms in human genes and variability in opioid analgesia and side effects among patients treated for moderate or severe pain. The evidence supporting a role of certain alleles, genotypes or haplotypes in modulation of opioid analgesia is derived from a limited number of studies, a limited number of genes and a limited number of opioids. Although several interesting candidates have emerged as potentially relevant factors, only for one polymorphism, the prevalent 118A>G of the micro-opioid receptor, the accumulated evidence is sufficient to suggest a clinically relevant effect for an opioid used for moderate or severe pain. Still the data are valid only at the group level and cannot be used to predict treatment outcome in individual patients. Only a few of the symptoms often seen as opioid adverse effects in palliative care, such as nausea, vomiting, constipation and sedation, have been associated with genetic variants in various genes, but the results have been based on case reports, healthy volunteers or post-operative patients. So far, there is no clear evidence that genetic markers can be used to predict opioid efficacy or adverse effects in palliative care patients. This reflects the general lack of studies performed in the context of palliative care, the lack of sufficiently scaled studies and the lack of international standards for the assessment of subjective symptoms.