PubMed/Medline, EMBASE, Web of Science, and Cochrane Library databases were comprehensively and systematically searched from inception to 31 December 2020, for studies published in English. The main search items were as follows: (“Menopause Hormone Therapy” OR “hormone therapy” OR “estrogen therapy” OR “estradiol therapy”) AND [“TC” OR “TG” OR “LDL” OR “HDL” OR “LP (a)” OR “lipid” AND (“postmenopausal women” OR “menopausal women” OR “menopause” OR “peri-menopausal women”). This review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement checklist (Moher et al., 2009), and that was registered on PROSPERO (Registration No. CRD42018092924). Two authors screened and evaluated all the abstracts and potentially eligible articles, any discrepancies between reviewers in the study selection were resolved via consultation with a third reviewer.

Articles that meet the following requirements were included: 1) original RCTs that were published in English; 2) administration of MTH for postmenopausal concerning factors, such as duration of therapy, route of administration, dosage, and types of regimens (E-alone or estrogen E + P); 3) inclusion of placebo, no treatment or non-MHT as a control group. For different regimens, regarding the effects of different types of estrogen on lipid profile, the same type of estrogen was required in 2 groups; 4) reporting the levels of TC, TG, LDL, HDL or Lp (a) as the outcome measures for lipid profile, and data were available directly from articles or could be calculated by mathematical formulas. The unit of TC, TG, LDL, and HDL was uniformly converted to mmol/L, and the unit of Lp(a) was converted to mg/L. As tibolone can alleviate menopause symptoms, studies that compared the effects of tibolone with MHT on lipid profile were included, while studies that concentrated on only the effects of tibolone were excluded from this review.

Data extraction of the studies included: 1) basic data of retrieved articles (title, the first author’s full name, year of publication, journal, etc.); 2)study design; 3) participants’ demographic characteristics (age, number of cases, etc.); 4)inclusion and exclusion criteria particularly for each article; 5) MHT-based data (name, dose, route of administration, the duration of treatment and type of regimen); 6) data related to control group (name, dose, route of administration, duration, type of regimen, etc.); 7) Serum lipid profiles. The data that provided baseline values and percentage changes after treatment only, which was unable to be converted into averages and standard deviations would be excluded. If raw data is needed, the corresponding author would be contacted to get more details. The Cochrane Risk of Bias check list (Higgins et al., 2011) was used to evaluate the risk of bias of randomized clinical trials.

Data analyzed was performed with the Cochrane Collaboration Review Manager (version 5.2) software, each outcome was expressed as mean ± standard deviation (SD). Heterogeneity among studies was estimated by I ² statistic. If I ² ≥ 50%, the random-effects model was used to perform the analysis; Otherwise, the fixed-effects model was utilized. We used the methods recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Ver. 6.2) to resolve the post-treatment data in some trials (Higgins et al., 2021). Millimoles per liter (mmol/L) will be used to measure TC, TG, LDL, and HDL while milligrams per liter (mg/L) were used to measure Lp(a).

Forty-seven studies (Cheng et al., 1993; Munk-Jensen et al., 1994; Samantray KV et al., 1994; Conard et al., 1995; Miller et al.,1995; Binder et al., 1996; Draper et al., 1996; Haines et al., 1996; Milner et al., 1996; Perrone et al., 1996; Conard et al., 1997; Heikkinen et al., 1997; Tuck et al., 1997; Espeland et al., 1998; Meschia et al., 1998; Lewis-Barned et al., 1999; Mijatovic et al., 1999; Davidson et al., 2000; Seed et al., 2000; Bunyavejchevin and Limpaphayom, 2001; Gräser et al., 2001; Luyer et al., 2001; Teede et al., 2001; Duvernoy et al., 2002; Samsioe et al., 2002; Ulloa et al., 2002; Wakatsuki et al., 2002; Yang et al., 2002; Hemelaar et al., 2003; Jirapinyo et al., 2003; Oral and Ozbaşar, 2003; Sanada et al., 2003; Wakatsuki et al., 2003; Stevenson et al., 2004; Bukowska et al., 2005; Zegura et al., 2006; Demirol et al., 2007; Fernandes et al., 2008; Villa et al., 2008; Ziaei et al., 2010; Cayan et al., 2011; Villa et al., 2011; Terauchi et al., 2012; Labos et al., 2013; Gregersen et al., 2019; Kotecha et al., 2020; Vaisar et al., 2021) compared the effects of MHT therapy and placebo on blood lipids. The duration of MHT was classified into the following periods: < 3 months, 3–5 months, 6–12 months, 13–24 months, and >24 months. For articles that evaluated the effects of MHT on lipid profile at multiple time points, the result in each time point was included as separate data.

The meta-analysis of data demonstrated that intake MHT could significantly reduce the serum TC (Miller et al., 1995; Binder et al., 1996; Bunyavejchevin and Limpaphayom, 2001; Cayan et al., 2011; Cheng et al., 1993; Conard et al., 1995; Conard et al., 1997; Davidson et al., 2000; Draper et al., 1996; Duvernoy et al., 2002; Fernandes et al., 2008; Gräser et al., 2001; Gregersen et al., 2019; Haines et al., 1996) (WMD: −0.43, 95% CI: −0.53 to −0.33, I ² = 93%) (Figure 4A) and LDL (Miller et al., 1995; Binder et al., 1996; Bunyavejchevin and Limpaphayom, 2001; Cayan et al., 2011; Cheng et al., 1993; Conard et al., 1995; Conard et al., 1997; Davidson et al., 2000; Draper et al., 1996; Duvernoy et al., 2002; Fernandes et al., 2008; Gräser et al., 2001; Gregersen et al., 2019; Haines et al., 1996) (WMD: −0.47, 95% CI: −0.55 to −0.40, I ² = 87%) throughout almost all treatment duration (Figure 4B). Except the duration between half year to 1 year (WMD: −0.08, 95% CI: −0.13 to −0.03), there was no significant difference in reducing TG (Cheng et al., 1993; Conard et al., 1995; Miller et al., 1995; Binder et al., 1996; Conard et al., 1997; Davidson et al., 2000; Bunyavejchevin and Limpaphayom, 2001; Bukowska et al., 2005; Cayan et al., 2011), (Duvernoy et al., 2002), (Haines et al., 1996; Heikkinen et al., 1997; Gräser et al., 2001; Fernandes et al., 2008; Gregersen et al., 2019) between the two groups (WMD: −0.00, 95% CI: −0.06 to 0.05, I ² = 84%) (Figure 4C). While come to Lp(a) (Bukowska et al., 2005; Conard et al., 1995; Conard et al., 1997; Davidson et al., 2000; Demirol et al., 2007; Espeland et al., 1998; Gregersen et al., 2019; Haines et al., 1996; Hemelaar et al., 2003; Kotecha et al., 2020; Meschia et al., 1998; Mijatovic et al., 1999; Milner et al., 1996; Samsioe et al., 2002), the results showed that MHT could remarkably decrease Lp(a) (WMD: −49.46, 95% CI: −64.27 to −34.64, I ² = 89%) (Figure 4E). However, the similar trend was only observed in periods of 6–12 months and >24 months. Data from 43 studies suggested an ignorable change in HDL (Miller et al., 1995; Binder et al., 1996; Bukowska et al., 2005; Bunyavejchevin and Limpaphayom, 2001; Cayan et al., 2011; Cheng et al., 1993; Conard et al., 1995; Conard et al., 1997; Davidson et al., 2000; Draper et al., 1996; Duvernoy et al., 2002; Fernandes et al., 2008; Gräser et al., 2001; Gregersen et al., 2019) (WMD: −0.00, 95% CI: −0.05to 0.05, I ² = 94%) (Figure 4D).

A total of 16 studies (Hemelaar et al., 2003; Bukowska et al., 2005; Meschia et al., 1998; Perrone et al., 1996; Seed et al., 2000; Zegura et al., 2006; Casanova et al., 2015; Casanova et al., 2009; Castelo-Branco et al., 2007; Faguer de Moustier et al., 1989; Hemelaar et al., 2006; Lahdenperä et al., 1996; Sendag et al., 2002; Taskinen et al., 1996; Tilly-Kiesi et al., 1996; Abbas et al., 2004) that enrolled 670 participants in oral MHT group and 676 in transdermal MHT group were analyzed. When comparing the effects between 2 groups, the result indicated that oral MHT could significantly decreased LDL-C (Hemelaar et al., 2003; Meschia et al., 1998; Perrone et al., 1996; Seed et al., 2000; Casanova et al., 2015; Zegura et al., 2006; Casanova et al., 2009; Castelo-Branco et al., 2007; Faguer de Moustier et al., 1989; Hemelaar et al., 2006; Lahdenperä et al., 1996; Sendag et al., 2002; Taskinen et al., 1996; Tilly-Kiesi et al., 1996) (WMD: 0.23, 95%CI: −0.31 to −0.14, I² = 28%) (Figure 5B) while there was no significant difference in TC (Hemelaar et al., 2003; Meschia et al., 1998; Perrone et al., 1996; Seed et al., 2000; Casanova et al., 2015; Casanova et al., 2009; Zegura et al., 2006; Castelo-Branco et al., 2007; Faguer de Moustier et al., 1989; Hemelaar et al., 2006; Lahdenperä et al., 1996; Sendag et al., 2002; Taskinen et al., 1996; Tilly-Kiesi et al., 1996) (WMD: −0.13, 95% CI: −0.30 to 0.04, I ² = 69%) (Figure 5A). However, the result revealed that oral MHT may significantly increase TG (Bukowska et al., 2005; Hemelaar et al., 2003; Perrone et al., 1996; Seed et al., 2000; Casanova et al., 2015; Casanova et al., 2009; Zegura et al., 2006; Castelo-Branco et al., 2007; Faguer de Moustier et al., 1989; Hemelaar et al., 2006; Lahdenperä et al., 1996; Sendag et al., 2002; Taskinen et al., 1996; Tilly-Kiesi et al., 1996) (WMD: 0.12, 95% CI: 0.04 to 0.21, I² = 50%) (Figure 5C), while both HDL (Bukowska et al., 2005; Hemelaar et al., 2003; Perrone et al., 1996; Seed et al., 2000; Casanova et al., 2015; Casanova et al., 2009; Zegura et al., 2006; Castelo-Branco et al., 2007; Faguer de Moustier et al., 1989; Hemelaar et al., 2006; Lahdenperä et al., 1996; Sendag et al., 2002; Taskinen et al., 1996; Abbas et al., 2004) (WMD: -0.02, 95% CI: −0.10 to 0.06, I ² = 84%) (Figure 5D) and Lp(a) (Meschia et al., 1998; Seed et al., 2000; Hemelaar et al., 2003; Bukowska et al., 2005; Hemelaar et al., 2006; Zegura et al., 2006) (WMD: 5.04, 95% CI: −20.32 to 30.41, I ² = 0%) had no significance (Figure 5E).

The studies were classified according to the dosage of estrogen. A total of 10 studies (Cheng et al., 1993; Stadberg et al., 1996; Taechakraichana et al., 2000; Sanada et al., 2003; Wakatsuki et al., 2003; de Kraker et al., 2004; Koh et al., 2004; Christodoulakos et al., 2006; Villa et al., 2008; Xue et al., 2016)that enrolled 584 participants in low-dose estrogen group and 594 in conventional dose estrogen group were analyzed. 1mg/day or less of Estradiol valerate or 17 β-estradiol, 0.3 mg/day or less of conjugated estrogens were defined as low dose estrogen.

The meta-analysis result showed that the low-dose estrogen led to a significant reduction in TG (Cheng et al., 1993; Sanada et al., 2003; Villa et al., 2008; Wakatsuki et al., 2003; Christodoulakos et al., 2006; Stadberg et al., 1996; Koh et al., 2004; Taechakraichana et al., 2000; Xue et al., 2016) (WMD: −0.18, 95% CI: −0.32 to −0.03, I ² = 93%) (Figure 6C) and HDL-C (Cheng et al., 1993; Sanada et al., 2003; Villa et al., 2008; Wakatsuki et al., 2003; Christodoulakos et al., 2006; de Kraker et al., 2004; Stadberg et al., 1996; Koh et al., 2004; Taechakraichana et al., 2000; Xue et al., 2016) (WMD: −0.05, 95% CI: −0.07 to −0.04, I ² = 36%) (Figure 6D) comparing with the conventional-dose estrogen. There was no significant on TC (Cheng et al., 1993; Sanada et al., 2003; Villa et al., 2008; Wakatsuki et al., 2003; Abbas et al., 2004; Christodoulakos et al., 2006; de Kraker et al., 2004; Stadberg et al., 1996; Koh et al., 2004; Taechakraichana et al., 2000) (WMD: −0.11, 95% CI: −0.26 to 0.04, I ² = 86%) (Figure 6A) and LDL-C (Cheng et al., 1993; Stadberg et al., 1996; Taechakraichana et al., 2000; Sanada et al., 2003; Wakatsuki et al., 2003; de Kraker et al., 2004; Koh et al., 2004; Christodoulakos et al., 2006; Villa et al., 2008; Xue et al., 2016) (WMD: 0.06, 95% CI: −0.17 to 0.29, I² = 96%) (Figure 6B). Because of only one study evaluated the effects of different doses on Lp(a), meta-analysis was not carried out.

As tibolone is widely used in mitigating the menopause symptoms, it is necessary to compare the effects of conventional MHT therapy with tibolone on lipids profile. A total of 13 studies (Cayan et al., 2011; Kotecha et al., 2020; Milner et al., 1996; Ziaei et al., 2010; Christodoulakos et al., 2006; Castelo-Branco et al., 1999; Farish et al., 1999; Koh et al., 2003; Koh et al., 2005; Mendoza et al., 2002; Pan et al., 2002; Siseles et al., 1995; Yang et al., 1999)that enrolled 646 participants in conventional MHT group and 828 in tibolone group were analyzed. The outcomes of meta-analysis presented the significantly increasing TG (Cayan et al., 2011; Kotecha et al., 2020; Milner et al., 1996; Ziaei et al., 2010; Christodoulakos et al., 2006; Castelo-Branco et al., 1999; Farish et al., 1999; Koh et al., 2003; Koh et al., 2005; Mendoza et al., 2002; Pan et al., 2002; Siseles et al., 1995; Yang et al., 1999) (WMD:0.42, 95%CI: 0.18 to 0.65, I² = 98%) (Figure 7C) and HDL-C (Cayan et al., 2011; Kotecha et al., 2020; Milner et al., 1996; Ziaei et al., 2010; Christodoulakos et al., 2006; Castelo-Branco et al., 1999; Farish et al., 1999; Koh et al., 2003; Koh et al., 2005; Mendoza et al., 2002; Pan et al., 2002; Siseles et al., 1995; Yang et al., 1999) (WMD: 0.36, 95% CI: 0.27 to 0.45, I² = 95%) (Figure 7D) concentration while significantly decreasing LDL-C (Cayan et al., 2011; Kotecha et al., 2020; Milner et al., 1996; Ziaei et al., 2010; Christodoulakos et al., 2006; Castelo-Branco et al., 1999; Farish et al., 1999; Koh et al., 2003; Koh et al., 2005; Mendoza et al., 2002; Pan et al., 2002; Siseles et al., 1995; Yang et al., 1999) (WMD: −0.35, 95% CI: −0.50 to −0.19, I² = 87%) (Figure 7B) concentration in conventional MHT group. No significant difference was identified in TC (Cayan et al., 2011; Kotecha et al., 2020; Milner et al., 1996; Christodoulakos et al., 2006; Castelo-Branco et al., 1999; Farish et al., 1999; Koh et al., 2003; Koh et al., 2005; Mendoza et al., 2002; Pan et al., 2002; Siseles et al., 1995; Yang et al., 1999) (WMD: 0.15, 95% CI: −0.15 to 0.44, I² = 96%) (Figure 7A) and Lp(a) (Milner et al., 1996; Farish et al., 1999; Demirol et al., 2007; Kotecha et al., 2020) (WMD: −18.31, 95% CI: −51.84 to 15.22, I² = 56%) (Figure 7E) concentration between two groups.

In total, 8 studies (Samantray KV et al., 1994; Miller et al., 1995; Farish et al., 1996; Wakatsuki and Sagara, 1996; Davidson et al., 2000; Hemelaar et al., 2003; Zegura et al., 2006; Fernandes et al., 2008) that enrolled 836 participants in E-alone group and 818 in E + P group met the criteria of eligibility. The micronized progesterone was used in 2 studies as separate group (Miller et al., 1995; Espeland et al., 1998) and synthetic progestogen was utilized in all these 8 studies.

The results revealed that E + P regimen significantly increased the concentration of TC (Miller et al., 1995; Davidson et al., 2000; Fernandes et al., 2008; Hemelaar et al., 2003; Samantray KV et al., 1994; Zegura et al., 2006; Farish et al., 1996; Wakatsuki and Sagara, 1996) (WMD: 0.15, 95% CI: 0.09 to 0.20, I ² = 18%) (Figure 8A), LDL-C (Miller et al., 1995; Fernandes et al., 2008; Hemelaar et al., 2003; Samantray KV et al., 1994; Zegura et al., 2006; Farish et al., 1996; Wakatsuki and Sagara, 1996) (WMD: 0.12, 95% CI: 0.07 to 0.17, I ² = 29%) (Figure 8B), HDL-C (Miller et al., 1995; Fernandes et al., 2008; Hemelaar et al., 2003; Samantray KV et al., 1994; Zegura et al., 2006; Farish et al., 1996; Wakatsuki and Sagara, 1996) (WMD: 0.10, 95% CI: 0.03 to 0.18, I ² = 87%) (Figure 8D), and Lp(a) (Farish et al., 1996; Espeland et al., 1998; Davidson et al., 2000; Hemelaar et al., 2003; Zegura et al., 2006) (WMD: 44.58, 95% CI:28.09 to 61.06, I² = 90%) (Figure 8E) concentration compared with E-alone. No significant difference was found in TG (Miller et al., 1995; Fernandes et al., 2008; Hemelaar et al., 2003; Samantray KV et al., 1994; Zegura et al., 2006; Farish et al., 1996; Wakatsuki and Sagara, 1996) concentration between these two groups (WMD: 0.05, 95% CI: −0.04 to 0.13, I² = 64%) (Figure 8C).

Considering that most of the pooled outcomes had an I² greater than 50%, one-by-one exclusion was performed as a sensitivity analysis to confirm the robustness of the outcomes. While omitting the study de Kraker 2004 (de Kraker et al., 2004), low-dose estrogen seems to decrease TC significantly (MD: −0.17,95% CI: −0.31 to −0.02) (Figure 9). The cause of unstable results may be attributed to the difference type of estrogen used in this study. Also, an unstable result was found in TG of comparing E-alone and E + P regimen. When study of writing−group 1995 (Miller et al., 1995) was excluded, E + P group could significantly higher TG (MD: 0.08, 95% CI: 0.01–0.15) (Figure 10) than Estrogen alone. The longer period of using MPA may be a source of instability. Egger test and funnel plots suggested that there was little indication of publication bias in studies with more than 10 trials (Figure 11).

Endogenous sex hormones are involved in the pathogenesis of cardiovascular disease (CVD) in women. Studies have shown that estradiol (E2), the major form of ovarian estrogen before menopause, plays an active role in metabolic actions (Franck et al., 2013). Higher estrone levels were related to a higher brachial flow-mediated dilation (ie, better endothelial function) (Thurston et al., 2018). After menopause there is a drastic change in the endogenous hormonal milieu, with a decrease in estradiol. And the circulating estrone (E1) levels are relatively higher than E2. E1 is produced mostly by the conversion of androgens in peripheral tissues, and could be also converted from E2 by 17 β⁃ Hydroxysteroid dehydrogenase, E1 secretion also decreased after menopause and was equivalent to nearly 1/3 before menopause (Qureshi et al., 2020). Studies showed that higher E1 is associated with more stable plaque (Cortés Yamnia et al., 2020) and better endothelial function (Thurston et al., 2018), lower levels of E1 have been associated with increased all-cause mortality among postmenopausal women (de Padua Mansur et al., 20122012), which proved the importance of estrogen on CVD. In addition to E1 and E2, sex hormone binding globulin (SHBG) and testosterone (T) may be associated with future risk of CVD also. One study showed that a more androgenic hormone profile (i.e., higher levels of free T and lower levels of SHBG) was associated with greater Coronary Artery Calcium (CAC) progression up to 10 years in postmenopausal women (Subramanya et al., 2019). In summary, as deficiency of endogenous estrogen after menopause and the importance of estrogen for CVD, the exogenous estrogen based MHT should be benefit for CVD and related high-risk factors in theory.

Our systematic review indicated that compared with placebo or no treatment, MHT could significantly decrease the concentrations of TC, LDL-C, and Lp (a). Lp(a) is an independent risk factor for CVD and recurrent ischemic stroke (Nordestgaard et al., 2010), the previous study showed the similar result of MHT on Lp(a) with us (van Dam-Nolen et al., 2021). As for the TG concentration, previous study had showed that MHT could significantly increase it (Stevenson et al., 2015). However, no significant difference in TG between two groups was found in our study. Hence, generally speaking, MHT was associated with favorable changes in lipid parameters whether short-term or long-term using in postmenopausal women.

The bioavailability of oral estrogen is mainly low due to first-pass metabolism, which may result in adverse reactions that influence the risk of CVD. Transdermal MHT is more appropriate for cases with a high-risk of CVD or dyslipidemia than oral agents. The results of our study showed that oral MHT significantly increased TG concentration compared with transdermal MHT. In addition, a meta-analysis conducted in 2006 revealed that oral MHT adversely affected C-reactive protein (CRP) level (Ambikairajah et al., 2019). Therefore, for women with hypertriglyceridemia or other high-risk factors of CVD, transdermal route is recommended. However, oral MHT is associated with positive effects in LDL-C concentration in our study. As we know, the LDL-C concentration is the main risk factor for the occurrence and development of atherosclerosis, and was regarded as an important index to assess the risk of atherosclerotic CVD (ASCVD) (Stone et al., 2013; Jacobson et al., 2015). Hence, for women without any risk of CVD or hypertriglyceridemia, oral MHT could possibly provide greater benefits.

Considering the safety factor, the minimum effective dose of estrogen was recommanded (Menopause Subgroup, Chinese Society of Obstetrics and Gynecology, Chinese Medical Association, 2018). However, whether the low-dose MHT could achieve the same effects on lipid profile as conventional-dose MHT is still confused. One study indicated that low-dose MHT was associated with higher levels of TC and LDL-C, lower TG level (Casanova et al., 2015). Our study showed the similar benefit on TG in low-dose MHT group, but no significant difference in TC and LDL-C levels between two groups. Furthermore, low-dose MHT was found could decrease HDL-C level. Epidemiologically, a low plasma level of HDL-C was associated with an increased risk of ischemic CVD (Haase et al., 2012). Taken together, the advantage of low-dose MHT on lipid profile was possibly only confined to the TG level.

Tibolone is a synthetic hormone with estrogenic, progestogenic, and androgenic properties, and was widely used for alleviating menopausal symptoms in postmenopausal women. Tibolone has shown promising effects on improving depression and libido, and does not increase breast density (Cummings et al., 2008). As for its effects on lipid profile, a meta-analysis10 conducted in 2017 concluded that there was no significant difference between conventional MHT and Tibolone in Lp(a) concentration, which is similar to our findings. While conventional MHT was found with lower LDL-C level and higher HDL-C level compared with Tibolone, while higher TG concentration. It is suggested that tibolone is more beneficial on TG concentration.

Progestogens are indicated as a part of systemic hormone therapy in women with an intact uterus, preventing estrogen-induced endometrial hyperplasia and cancer during estrogen exposure. However, an increased risk of CHD in women receiving estrogen plus progestogen therapy rather than in those receiving CEE alone was reported (Falkeborn et al., 1992). Thus, it should be indicated whether progestogen contributes to adverse outcomes of CVD. However, no large-scale RCT has evaluated the lipid profile according to the type of progestogen used. A previous observational study revealed that the addition of progestogens blunts the lipid-related effects (Shufelt and Manson, 2021), and a meta-analysis performed in 2017 indicated that there was no significant difference in the reduction of Lp(a) concentration by E-alone compared with E + P (Anagnostis et al., 2017). The results in our study showed that E + P regimen weakened the benefits of estrogen mono-therapy. However, it should be noted that the progestogens included in our analysis were mainly composed of synthetic progestogen, and further research is required to explore whether natural progesterone could positively influence lipid profile.

Except for routine MHT, selective estrogen receptor modulators (SERMs), such as tamoxifen and raloxifene, are widely used for patients with breast cancer or osteoporosis. SERMs mimic estrogen action in certain tissues while opposing it in others. The effect of SERMs on lipids profile is also an issue worthy of attention. The meta-analysis had showed that tamoxifen can alter the lipid profile in females, particularly by decreasing TC, LDL-C and HDL–C (Alomar et al., 2022). Rraloxifene can increase HDL-C and lower LDL-C and TC (Yang et al., 2021). Thus, SERMs is beneficial to blood lipids in general.

In addition, although the result showed the positive effects of MHT on lipid profile, it needs to be emphasized that MHT is not recommended as first-line therapy for dyslipidemia or for reducing the risk of cardiovascular disease (Panagiotis et al., 2020). For postmenopausal women with carotid atherosclerosis, the prospective study had showed that total estradiol was associated with presence of vulnerable carotid plaque as well as increased risk of stroke (Glisic et al., 2018). Therefore, it is recommended to start MHT in women <60 years of age or <10 years since menopause for the beneficial effects on CVD outcomes (2019 Surveillance of Menopause, 2019; El Khoudary et al., 2020).

For dyslipidemia, the most commonly used medication is HMG-CoA reductase inhibitors (ie, statins). Statin therapy can also have effects on gonada steroidogenesis, since this process requires cholesterol as a biochemical substrate. LDL-C has been shown to be a preferential precursor for the production of ovarian steroid hormones (Grummer and Carroll, 1988). However, no reduction in E2 or E1 in postmenopausal women taking statins, despite a significant decrease in their LDL-C levels (Bairey Merz et al., 2002). But there are many studies showing an association between statin treatment and a reduction in testosterone levels (Stamerra et al., 2021). For polycystic ovary syndrome (PCOS) women, statins could decrease testosterone and Luteinizing hormone (LH)/Follicle stimulating hormone (FSH) ratio (Seyam et al., 2017), which is beneficial in treatment of PCOS. However, the role of statins for primary prevention in postmenopausal women is debated (Cangemi et al., 2017). Evidence-based data of statins for the reduction of CVD events and all-cause mortality in primary prevention in postmenopausal women is needed (El Khoudary et al., 2020).

The limitations of the present study should be pointed out. Firstly, among the eligible studies, few studies were specifically designed to evaluate the effects of MHT on lipid profile as the primary outcome, restricting the generalization of our findings. Secondly, the lipid profile at baseline in the majority of the included studies was almost normal, while it remained elusive whether MHT would have the similar effects on lipid profile in women with dyslipidemia. Thirdly, owing to the small sample size, the comparison between the effects of different types of progestogen on lipid profile was not comprehensively performed. Therefore, further research needs to be conducted to eliminate the above-mentioned limitations and to confirm our findings.