Table 2. Characteristics of studies excluded from meta-analysis of treatment of premenstrual syndrome

Study Participants Intervention progesterone Reason for exclusion Reported results Side effects Comments

Richter et al, 1984²⁴ 40 women referred from general practice 400 mg/day progesterone suppository luteal phase No prediagnosis of PMS; women recruited with self diagnosis More women believed that progesterone gave symptom relief than placebo No withdrawals due to side effects No difference in improvement seen between treatment groups

Smith et al, 1975²⁹ 14 randomised 50 mg intramuscular progesterone every other day in luteal phase Insufficient published results for data analysis 3 women felt better on progesterone, 3 women felt better during progesterone free months; 8 women found no difference No withdrawal, side effects not mentioned Trial was a crossover of 4 treatment regimens: progesterone; progesterone plus spironolactone; spironolactone; placebo injections and tablets

Sampson, 1979²⁷ 32 randomised, 24 completed crossover 2x200 mg/day, 2x400 mg/day suppositories luteal phase for 2 months Low Jadad score No significant difference from placebo 400 mg: 7 drop outs; 800 mg: 9 drop outs; none due to side effects

Baker et al, 1995²⁶ 17 completed multiple crossover 2x200 mg/day vaginal suppositories luteal phase for 7 months Low Jadad score No overall difference from placebo; significant improvement for tension, mood swings irritability, control None reported Trial assessed only psychological symptoms

Coppen, 1969³³ 17 completed parallel trial 2x7.5 mg/day norethisterone on day 16-25 of cycle Data presented not suitable for extraction Not effective in improving premenstrual symptoms None reported Norethisterone was also compared with diuretic, Dytide

Hoffmann, 1988³⁴ 161 completed parallel trial 2x10 mg/day dydrogesterone on day 12-menses for 3 cycles No prospective diagnosis of PMS No clinically relevant effect 38 drop outs, 3 due to side effects

Haspels, 1980³⁵ 123 completed parallel trial 2x10 mg/day dydrogesterone on day 12-menses for 4 cycles Subgroup of patients from included study[15] Significantly better than placebo for psychological symptoms and clinically better for somatic symptoms 27 drop outs, none due to side effects British arm of European study

Jordheim, 1972³⁶ 35 completed parallel trial 3x2.5 mg medroxyprogesterone 10 days before menstruation No placebo arm, no prospective diagnosis of PMS No significant effect None stated Medroxyprogesterone compared with medroxyprogesterone plus diuretic

Kerr, 1980³⁷ 67 completed 2x10 mg/day dydrogesterone on day 12-menses for 4 cycles No preliminary diagnosis; single blind trial Unclear: "dydrogesterone is a useful agent" 36 drop outs, 3 due to side effects Trial funded by pharmaceutical company

Hellberg, 1991³⁸ 38 completed crossover 5 mg/day medroxyprogesterone acetate for 3 cycles No prospective diagnosis of PMS Significantly better than placebo 5 drop outs, none due to side effects 2 interventions compared with placebo; spironolactone (50mg/day) also better than placebo

Sampson, 1988³⁹ 69 completed crossover 2x10 mg/day dydrogesterone for 14 days of cycle for 4 cycles Data presented not suitable for extraction Significant decrease in pain with menstrual bleeding and breast symptoms only 39 drop outs, 5 due to side effects

Sampson, 1982⁴⁰ Same patient group as above

Taylor, 1977⁴² 50 completed 2x10 mg/day dydrogesterone on day 12-26 of cycle for 2 or more cycles Open trial; no prospective diagnosis Measureable improvement in 70% of patients None

Strecker, 1981⁴³ 31 completed 20 mg/day dydrogesterone on day 15-25 of cycle Open trial; no prospective diagnosis Beneficial for relief of some symptoms No side effects reported

Strecker, 1980⁴⁴ Same patient group as above

Morse, 1991⁴¹ 14 completed 20 mg/day dydrogesterone on day 17-27 for 3 cycles Open trial Some short term symptom relief No drop outs due to side effects Dydrogesterone v cognitive therapy and relaxation therapy