Table 1.
GRADE Quality of Evidence Assessment
| Study | Study Design | Risk of Bias | Imprecision | Inconsistency | Indirectness | Publication Bias | Comments | Quality |
|---|---|---|---|---|---|---|---|---|
| Cachrimanidou 1993 [7] | RCT | Low | Not serious | Not serious | Not serious | Not serious | N/A | High |
| Brandes 2006 [5], Misakian 2003 [8] | Cross sectional | Low | Not serious | Not serious | Not serious | Not serious | Both studies used data from Women's Health Study | High |
| Amir 2005 [9] | Cross sectional | Low | Serious | Serious | Not serious | Not serious | Potential confounding from participants being treated with IVF. Inconsistency with other studies due to conclusion that GnRH therapy increased risk of migraine. Potential imprecision due to to small sample size. | Moderate |
| Calhoun 2008 [4] | Prospective cohort | Some | Serious | Serious | Not serious | Not serious | Potential confounding from participants having a history of medically intractable menstrual migraine. Inconsistency with other studies and imprecision due to 36% of participants in treatment group having no statistically significant change in outcome and being excluded from treatment group in final analysis. | Moderate |
| de Lignieres 1986 [10] | Double blind placebo-controlled crossover | Some | Serious | Not serious | Not serious | Not serious | Older age of participants and small sample size | Moderate |
| Facchinetti 2002 [11], Martin 2003 [12] | RCT | Some | Serious | Not serious | Not serious | Not serious | Older age of participants and small sample size | Moderate |
| MacGregor 2006 [13] | Double blind placebo-controlled crossover | Some | Serious | Not serious | Not serious | Not serious | Nonspecific participant selection criteria. Imprescision due to wide confidence intervals and small sample size | Moderate |
| Calhoun 2004 [14] | Open-label clinical trial | Some | Serious | Not serious | Not serious | Not serious | Lack of control and blinding. Imprecision due to small sample size and lack of control to compare results | Low |
| Johannes 1995 [15] | Self-reported 4-month diary | High | Serious | Not serious | Not serious | Not serious | Imprecision due to small sample size and lack of control to compare results | Low |
| Lichten 1995 [16] | Crossover | Some | Very serious | Not serious | Serious | Not serious | Potential confounding from participants having a history of medically intractable menstrual migraine. Indirectness present due to results not discussing the difference in response to treatment between participants, which is what the study design was for. Additionally, only women women that initially responded to treatment were progressed in the study. Considerable imprecision from small sample size and lack of consistent statistical analysis. | Low |
| Mattsson 2003 [17] | Survey | High | Very serious | Not serious | Not serious | Not serious | Increased risk of bias and imprecision due to self-reporting study design | Low |
| Pringsheim 2004 [6] | Survey | High | Very serious | Not serious | Not serious | Not serious | Increased risk of bias and imprecision due to self-reporting study design and lack of control group. | Low |
| Stewart 2000 [18] | Self-reported 98 day diary | High | Very serious | Not serious | Not serious | Not serious | Increased risk of bias and imprecision due to self-reporting study design and lack of control group. | Low |
| Lichten 1996 [19] | Open-label clinical trial | High | Very serious | Not serious | Not serious | Not serious | Lack of control group and single time administration. Imprecision due to small sample size and lack of control to compare results | Very low |
| Marcus 1999 [20], Murray 1997 [21] | Prospective cohort | High | Very serious | Serious | Not serious | Not serious | Very small sample sizes and potential confounding variables of pregnancy and / or hormone replacement therapy. Inconsistency due to conclusion that GnRH therapy decreased headache index. Imprecision due to small sample size and lack of control to compare results | Very low |
| Lichten 1996 [19] | Open-label clinical trial | High | Very serious | Not serious | Serious | Not serious | Small sample size and potential confounding from HRT use. Indirectness due to conclusions focusing on specific estrogen serum levels that precipitate migraine while study data focused on effect of timing in relation to withdrawal of estrogen on precipitation of migraine. Imprecision due to small sample size and lack of consistent statistical analysis | Very low |
| Somerville 1972 [22] | Crossover (case series) | High | Very serious | Not serious | Not serious | Serious | Small sample size and lack of statistical analysis. Potential risk of publication bias due to positive study results and being one of the first studies published on the topic. However this study is considered a landmark study in the field of hormonal migraines. | Very low |