Several classes of medication have been studied in the role of appetite stimulants. The best established are the progestogens - megestrol acetate and medroxyprogesterone acetate, and corticosteroids. Their mechanism of action is not well understood.
This is an active area of clinical research, and a number of other agents are also being investigated, including metoclopramide, dronabinol, EPA (fish oil), erthythropoietin, thalidomide, ghrelin, interferon, melatonin, non-steroidal anti-inflammatories, nandrolone, antidepressants such as mirtazepine, and atypical antipsychotics such as olanzapine.
What is known
There is evidence that megestrol acetate (MA) improves appetite (NNT = 4) and has a small effect on weight gain (NNT = 12) in advanced cancer, AIDS, and some other conditions, with the effect increasing at higher doses.  These benefits were seen when compared with placebo, but not in studies where it was compared with other drugs. Megestrol acetate has not been shown to improve quality of has a range if adverse effects. Thromboembolic complications resulted in death in one in 23 patients taking this medication.
A smaller number of studies have investigated medroxyprogesterone acetate (MPA). 
The duration of most studies was six to twelve weeks. It was not possible to draw a conclusion about the impact on patients’ quality of life, or what dose should be used.
Dexamethasone, prednisolone and methylprednisolone have all been shown to improve appetite in patients with cancer, and some studies have also shown increases in quality of life and wellbeing. There is not enough evidence to recommend doses or duration of treatment, and it is suggested that the effects tend to last four to eight weeks, while side effects increase with duration of treatment. [2,3]
Hydrazine sulphate has been shown to be ineffective in numerous studies, and should not be used as an appetite stimulant. 
Despite promising results in early studies, eicosapentaenoic acid (EPA) and cannabinoids have not gone onto produce clinically useful weight gain or improved appetite in phase III trials. [2,4] Side effects of EPA are mild but they affect quality of life, limiting its acceptability to patients. 
A systematic review of use of thalidomide to treat cachexia found few high quality studies, and no recommendation could be made based on the available evidence. 
Non-steroidal anti-inflammatory drugs have been investigated as a possible treatment for cachexia, based on their potential to modulate the pro-inflammatory components of the cachexia anorexia syndrome. Two systematic reviews have evaluated the literature on use of non-steroidal anti-inflammatory drugs, and suggest that whilst some improvements in weight and physical performance have been shown, the evidence is of poor quality and NSAIDs cannot be recommended to treat cachexia except in clinical trials at this stage. [7,8]
What it means in practice
- Stimulating appetite often does not reverse cachexia. 
- Progestogens or corticosteroids can be used to improve appetite in palliative care patients. The benefits in terms of quality of life and well-being have not been established, and clinicians should assess these regularly in individual patients being treated. There are no recommendations to guide the choice of dose or the duration of treatment. 
- Choice of drug should take into account the side effect profile, cost and burden of treatment. [1,9]
- It is recommended that other agents only be used in the setting of a clinical trial.