Online Submissions: http://www.journaltcm.

com J Tradit Chin Med 2020 June 15; 40(3): 343-354

info@journaltcm.com ISSN 0255-2922
© 2020 JTCM. All rights reserved.

SYSTEMATIC REVIEW

Effectiveness of tonifying-kidney and regulating-liver therapy on di-

minished ovarian reserve: a systematic review and Meta-analysis of
randomized controlled trials

Liu Liuqing, Liu Yanfeng, Yang Ming, Xu Guiqin, Li Ruiqi, Xu Xiuli, Pan Xue, Liang Jialing
aa
Liu Liuqing, Xu Xiuli, Pan Xue, Liang Jialing, Department METHODS: The literature was comprehensively
of Gynecology, the 1st Clinical Medical College of Beijing searched up to August 2019 using four Chinese
University of Chinese Medicine, Beijing 100700, China and three English electronic databases to extract
Liu Yanfeng, Department of Gynecology, Dongzhimen Hos-
randomized clinical trials (RCTs) comparing Tradi-
pital of Beijing University of Chinese Medicine, Beijing
100700, China
tional Chinese Medicine tonifying-kidney and regu-
Yang Ming, Center for Evidence Based Chinese Medicine, lating-liver prescriptions (combined with hormone
Beijing University of Chinese Medicine, Beijing 100029, therapy or not) with Western Medicine. Data quali-
China ty evaluation was conducted using the Cochrane
Xu Guiqin, Department of Scientific Research Management, risk of bias tool. Meta-analysis was conducted us-
the 1st Affiliated Hospital of Anhui University of Chinese ing Revman 5.3 software with effect estimates pre-
Medicine, Hefei 230001, China sented as mean difference (MD), risk ratio (RR), and
Li Ruiqi, Department of Gynecology, Shunyi Hospital, Bei- 95% confidence interval (CI).
jing Hospital of Traditional Chinese Medicine, Beijing
101300, China RESULTS: A total of nine RCTs with 512 partici-
Supported by the Fundamental Research Funds for the pants were extracted and eligible for Meta-analysis.
Central Universities: Study on the Mechanism of Tonify- There were no significant differences between Chi-
ing-Kidney and Smoothing-Liver on Rats With Diminished nese medicine and Western Medicine on basal se-
Ovarian Reserve Based on PI3K-AKT-mTOR Signaling Path-
rum follicle-stimulating hormone (FSH) level (MD
way (No. 2019-JYB-XS-148); National Natural Science Foun-
0.11, 95% CI －0.52 to 0.74, 392 participants, seven
dation of China: Study on the Mechanism of Regulating En-
dometrial Receptivity by Erbu Zhuyu Decoction Based on trials), anti-Müllerian hormone level (MD 0.48, 95%
NK/M-CSF/NO Signaling Pathway (No. 81473721) CI － 0.62 to 1.58, 95 participants, two trials), and
Correspondence to: Prof. Liu Yanfeng, Department of Gy- the FSH and luteinizing hormone ratio (MD 0.01,
necology, Dongzhimen Hospital of Beijing University of Chi- 95% CI －0.95 to 0.96, 115 participants, two trials).
nese Medicine, Beijing 100700, China. liuyaf8888@sina.com; Chinese medicine was more effective at improving
Liu Liuqing, Department of Gynecology, the 1st Clinical Traditional Chinese Medicine symptom scores (TC-
Medical College of Beijing University of Chinese Medicine, MSS) (MD －2.39, 95% CI －3.83 to －0.94, 160 par-
Beijing 100700, China. ahhfllq@163.com
ticipants, three trials), effective rate of TCMSS (RR
Telephone: +86-13910024921; +86-13466369901
1.18, 95% CI 1.02 to 1.36, 160 participants, three tri-
Accepted: November 20, 2019
als), antral follicle count (AFC) (MD 0.55, 95% CI 0.05
to 1.04, 155 participants, three trials), and FSH lev-
els at 3 months post-treatment (MD − 4.77, 95% CI
Abstract －6.09 to －3.45, 137 participants, two trials).
OBJECTIVE: To evaluate the effectiveness of the CONCLUSION: Compared with Western Medicine,
Traditional Chinese Medicine tonifying-kidney and tonifying-kidney and regulating-liver therapy is
regulating-liver therapy on diminished ovarian re- more effective at relieving symptoms and improv-
serve (DOR). ing AFC and FSH at 3 months post-treatment.

JTCM | www. journaltcm. com 343 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

© 2020 JTCM. All rights reserved. is.5,9,10 Age is another risk factor, whereas late menarche
(menarche age > 13 years) is a protective factor for the
Keywords: Ovarian reserve; Tonifying kidney and decline of ovarian reserve function.9,11
regulating liver; Randomized controlled trial; Sys- There is currently no definitive and effective method
tematic review; Meta-analysis to restore ovarian function in patients with DOR.5 In
addition to lifestyle changes and the avoidance of harm-
ful factors, hormonal replacement therapy (HRT),
INTRODUCTION which includes sequential therapy with estrogen and
progesterone (artificial menstrual cycle) and combined
Diminished ovarian reserve (DOR) is a normal physio- estrogen/progesterone, is a primary treatment for
logic occurrence for older women because of the age-re- DOR patients to correct low estrogen status, prevent
lated decline in ovarian function. However, many wom- the reproductive organs from shrinking, maintain regu-
en experience DOR much earlier. This is termed patho- lar menstruation, and prevention related complica-
logic DOR,1 and describes women of reproductive age tions. The androgen therapy of dehydroepiandros-
with normal menstruation who have a more dimin- terone supplementation can also improve ovarian func-
ished response to ovarian stimulation or fertility than tion, lower miscarriage rate, reduce aneuploidy, and in-
comparable women.2 crease pregnancy chances.12 Various ovulation induc-
Pathologic DOR may lead to decreased fertility and tion treatments and assisted reproductive techniques
poor pregnancy outcomes, such as high rates of preg- are also available for infertile DOR patients who wish
nancy loss.3 Abnormal ovarian reserve is also closely re- to become pregnant. However, although it is fast and
lated to poor ovarian response, according to the Bolo- effective, HRT has some contraindications.13 Assisted
gna criteria published by the European Society of Hu- reproductive technology also has a high risk of failure
man Reproduction and Embryology,4 and may be the and is expensive.14 The recovery of ovarian function
prodromal stage of ovarian failure.5 DOR can also ad- and the treatment of infertility caused by DOR there-
versely affect the outcome of in vitro fertilization-em- fore remains a substantial challenge.
bryo transfer (IVF-ET) to some extent.6 According to However, Traditional Chinese Medicine (TCM) has a
1995-1998 data from the United States, 10.55% of pa- good therapeutic effect on DOR. TCM has an overall
tients (1034/9802) in one infertility clinic were diag- effect on regulation and restoration of ovarian func-
nosed with DOR.3 Another study demonstrated that tion, as it can regulate endocrinological factors and pro-
DOR prevalence increased from 19% to 26% from tect ovarian function from damage.14 Meta-analysis in-
2004 to 2011 among 181 536 fresh, autologous assis- dicates that TCM may provide an effective and safe al-
tive reproductive technology (ART) cycles reported to ternative therapy for patients with DOR.15 From a
the Society for Assisted Reproductive Technology Clin- TCM perspective, DOR is categorized under menoxe-
ic Outcomes Reporting System.7 Thus, early medical nia, metrorrhagia and metrostaxis, infertility, etc. The
intervention for DOR is urgently needed to reverse de- most basic type of DOR is the TCM kidney deficiency
clining ovarian function and promote natural pregnan- pattern. The substantial mental stress that arises from
cy or increase the IVF-ET pregnancy rate. the important social roles women play in modern soci-
In 2015, the Practice Committee of the American Soci- ety means that liver depression is a common pattern.
ety for Reproductive Medicine noted the lack of a "uni- Some research indicates that kidney deficiency and liv-
formly accepted definition of DOR,"2 which makes the er depression is a common pattern,16,17 so tonifying-kid-
criteria for DOR difficult to determine. DOR is cur- ney and regulating-liver therapy is frequently used.
rently clinically diagnosed by an abnormal ovarian re- TCM prescriptions for this therapy consist mainly of
serve test4 comprising elevated but not menopausal bas- kidney-tonifying medicine (which nourishes kidney
al follicle-stimulating hormone (FSH), low anti-Mülle- Yin and kidney essence, and reinforces kidney Yang
rian hormone (AMH), low antral follicle count (AFC), and kidney Qi) and liver-regulating medicine (which
and other less frequently used markers (like FSH/LH smooths liver Qi, pacifies liver Yang, emolliates the liv-
[luteinizing hormone] ratio, estradiol level, and previ- er, and clears liver heat). These medicines replenish the
ous IVF outcome) among women with regular men- essence and blood and regulate the movement of Qi
strual periods.1,8 and blood, which restores the reproductive function of
DOR can be induced by multiple causes, which re- DOR patients. In this study, we aimed to evaluate the
main obscure. In addition to some idiopathic causes, role that kidney-tonifying and liver-regulating therapy
impaired ovarian reserve may be associated with genet- plays in DOR treatment using a systematic review and
ic factors, autoimmune factors, iatrogenic factors (such Meta-analysis.
as radiotherapy, chemotherapy, or surgery), and psycho-
logical and environmental factors.5,9 Underlying mecha-
nisms include excessive follicle atresia, ovulation dys- METHODS
function, direct destruction of organs or tissues, and Inclusion criteria
disturbances in the hypothalamic-pituitary-ovarian ax- Included studies had to meet four criteria. First, only

JTCM | www. journaltcm. com 344 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

randomized controlled trials (RCTs) reporting the ef- ed were basal serum FSH levels, AMH, and AFC.
fectiveness of tonifying-kidney and regulating-liver TCM symptom score (TCMSS), effective rate of TC-
therapy were eligible for inclusion (regardless of blind- MSS, FSH/LH ratio, and basal serum FSH tested at 3
ing). Second, eligible participants had to meet the months post-treatment were selected as additional out-
DOR criteria. As mentioned above, there are currently come indicators.
no uniform diagnostic criteria for DOR. According to
the Bologna criteria,4 AMH and AFC are two impor- Data sources and search strategies
tant indicators of ovarian reserve. One study of DOR Two authors (Liu LQ and Li RQ) were in charge of the
diagnostic criteria based on a PubMed search summa- data search. An extensive literature search was conduct-
rized 14 studies that defined DOR and showed that se- ed to identify potentially eligible studies in the follow-
rum FSH, AFC, and AMH levels were the most fre- ing electronic databases: China National Knowledge In-
quently evaluated indicators, and were thus most com- frastructure (CNKI), WanFang Database, Chinese VIP
monly used in diagnosis.8 On the basis of the literature Information, SinoMed, PubMed, Cochrane Library,
and these criteria, combined with the definition of and EMBASE. The search strategy used the following
DOR in the Chinese Expert Consensus about the Di- terms: "diminished ovarian reserve" and "(Traditional
agnosis And Treatment of Premature Ovarian Insuffi- Chinese Medicine) or (Chinese medicine) or (herbs) or
ciency,18 we defined DOR as an abnormal ovarian re- (complementary or alternative medicine)" and "ran-
serve test result in patients aged < 40 years: AFC (moni- domized." Records were searched up to August 2019
tored by ultrasonography) < 5, or AMH < 1.1 ng/mL, and no language restriction was applied. For instance,
or 10 IU/L < basal FSH < 40 IU/L, or basal FSH/ the search strategy employed for PubMed is shown be-
basal LH > 3 (basal FSH and basal LH should be low.
tested on menstruation cycle days 2-4). Third, regard- #1 diminished ovarian reserve [Title/Abstract] OR di-
ing interventions, the experimental group had to be minished ovarian reserve [MeSH Terms]
treated with an oral TCM decoction comprising #2 decreased ovarian reserve [Title/Abstract] OR de-
mainly kidney-tonifying drugs (including kidney creased ovarian reserve [MeSH Terms]
Yin-nourishing drugs, kidney essence-nourishing #3 ovarian dysfunction [Title/Abstract] OR ovarian
drugs, kidney Yang-reinforcing drugs, and kidney dysfunction [MeSH Terms]
Qi-reinforcing drugs) and liver-regulating drugs (liv- #4 #1 OR #2 OR #3
er-smoothing drugs, liver-emolliating drugs, liver-paci- #5 Traditional Chinese Medicine [Title/Abstract] OR
fying drugs, and liver heat-clearing drugs) according to Traditional Chinese Medicine [MeSH Terms]
the Pharmacopoeia of the People's Republic of China #6 Chinese medicine [Title/Abstract] OR Chinese
(2015)19 and the Traditional Chinese Pharmacy text- medicine [MeSH Terms]
book,20 combined or not with common Western Medi- #7 herbs [Title/Abstract]) OR herbs [MeSH Terms]
cine (WM) treatments (the presence of herbs with #8 complementary or alternative medicine [Title/Ab-
other efficacies in prescription or modified prescrip- stract] OR complementary or alternative medicine
tions were allowed). The control group had to be [MeSH Terms]
treated with HRT, androgen therapy, ovulation induc- #9 #5 OR #6 OR#7 OR #8
tion treatment, or ART. Finally, included studies had #10 randomized [All fields]
to report the results of FSH, or AFC, or AMH, #11 #4 AND #9 AND #10
which we designated as the main outcome indicators Selection of studies
owing to the important role they play in the diagno- Two reviewers (Liu Liuqing and Pan Xue) selected stud-
sis of DOR. ies independently. Records obtained were first checked
Exclusion criteria to eliminate duplicates. Then preliminary screening
Studies were excluded if they met the following condi- was conducted by reading the title and abstract to ex-
tions: (a) the sum number of kidney-tonifying drugs clude studies unrelated to the clinical treatment of
and liver-regulating drugs was less than 60% of the to- DOR by TCM. The full text of the retained records
tal number of drugs in the TCM prescription (includ- was then downloaded and read to select eligible studies
ing drugs added circumstantially), or the number of according to the inclusion and exclusion criteria. If the
kidney-tonifying medicines was less than two, or the full text of a study could not be obtained, we contacted
number of liver-regulating drugs was less than two; (b) the author to request the relevant content. If the au-
the therapeutic regimen details were not clearly report- thor did not reply or did not provide the requested con-
ed; (c) FSH, AMH, or AFC results were not included tent, the study was excluded. Any disagreement in this
in the outcome indicators; (d) data were incorrect or process was resolved with assistance from a third party
prescriptions were incomplete. (Liu YF).
Outcome measurements Data extraction
Following the commonly used test indexes reported in Two authors (Liu Liuqing and Xu Xiuli) screened the
the literature,8 the main outcome indicators we select- included studies independently and extracted relevant

JTCM | www. journaltcm. com 345 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

data. Excel 2016TM (Microsoft, Redmond, WA, USA) cles were identified. After eliminating duplicates, 554
was used to create a data abstraction form to record in- articles remained. We read the abstracts and the full
formation. The main contents of the data abstraction text of these and retained 69 studies according to the
form were: (a) general information about studies (title, inclusion criteria. After eliminating studies meeting the
year, author names, and publication information); (b) exclusion criteria, nine RCTs23-31 with 512 participants
methodological details (random sequence generation, were included. Figure 1 shows the study selection
allocation concealment, blinding); (c) participant infor- process.
mation (number of participants, diagnostic criteria);
(d) intervention information (therapeutic regimen, Description of studies
drugs or herbs, doses, and treatment course); (e) out- All nine RCTs were conducted with Chinese subjects
come information (FSH, AMH, AFC, TCMSS, effec- and involved 512 participants (sample sizes varied
tive rate of TCMSS, FSH/LH, FSH at 3 months from 40 to 70). The therapy group interventions in-
post-treatment, and adverse reactions). Any disagree- cluded TCM decoctions containing kidney-reinforcing
ment was resolved by discussion, with assistance from a and liver-regulating drugs, and a combination of TCM
third party (Yang M) if necessary. decoction and standard WM. The control groups were
given standard WM therapy. The treatment durations
Assessment of risk of bias in included studies were 3 months or three menstrual cycles. Eight studies
Two reviewers (Liu Liuqing and Yang Ming) indepen- compared TCM decoction with WM and one study
dently assessed the quality of included studies using tested TCM decoction combined with WM versus
the Cochrane Handbook risk of bias assessment tool. WM. Details of the characteristics of the included stud-
This tool comprises seven aspects: random sequence ies are shown in Table 1.
generation, allocation concealment, blinding of partici-
pants and personnel, blinding of outcome assessment, Risk of bias in included studies
incomplete outcome data, selective reporting data, and We evaluated the included studies from seven aspects
other bias. Each potential source of bias was graded as (Figure 2). All were RCTs, but only three24,29,31 reported
high, low, or unclear.21,22 Any disagreement was re- the randomization procedure (use of a random number
solved by discussion, with assistance from a third party table). Therefore, these three studies were classed as
(Xu Guiqin) if necessary. "low risk of bias"and the other six studies classed as
"unclear risk of bias." None of the nine studies men-
Statistical analysis
tioned allocation concealment so they were classed as
The Meta-analysis was conducted using Revman 5.3 having an "unclear risk of bias." None of the included
(Copenhagen, DK, the Nordic Cochrane Centre, Co-
studies mentioned blinding of participants and person-
chrane Collaboration, 2014) to combine more than
nel. In consideration of the obvious difference in dos-
one trial to estimate pooled intervention effects. Di-
age forms between the two patient interventions, we
chotomous data were summarized using risk ratios classed all the studies as "high risk of bias" for this as-
(RRs) with 95% confidence intervals (95% CIs). Con-
pect. None of the studies mentioned blinding of assess-
tinuous data were expressed as mean differences (MDs)
ment, so all were classed as "unclear risk of bias" for
with 95% CIs. Statistical heterogeneity was tested us-
this aspect. Regarding incomplete outcome data, one
ing the χ2 test and examining I 2, and data were com- study28 was classed as "high risk of bias" because of the
bined in a fixed-effect model if statistical heterogeneity
differences in sample size between different outcomes
was low. When P < 0.1 or I ² > 50%, which indicated
without explanation. Six studies were classed as "un-
the possibility of statistical heterogeneity, a random-ef-
clear risk of bias" because of the inconsistent number
fect model was used. In the event that a meta-analysis of participants before and after grouping (which did
was not possible, we planned to use a narrative analysis
not significantly influence the data analysis)26 or the ab-
of the results from individual studies. Subgroup analy-
sence of loss to follow-up reporting.24,25,29-31 Another
sis was conducted to determine the evidence for a dif-
two studies23,27 were classed as "low risk of bias." Re-
ferent intervention effect (only TCM treatment or garding selective reporting, one study25 did not report
TCM treatment combined with WM). A funnel plot
the outcomes to be observed in the methods, so we
was used to detect publication bias.
had to class it as "unclear risk of bias." The other eight
A protocol of study methods was registered in PROS-
studies were classed as "low risk of bias" for this aspect,
PERO and published on the website (https://www.crd.
as the outcomes to be observed that were mentioned in
york.ac.uk/prospero/). Trial registration number in
the methods were all reported in the results. Regarding
PROSPERO: CRD42019120832.
other forms of bias, six studies23,24,26-29 were academic dis-
sertations and one study31 was supported by the Shang-
RESULTS hai Municipal Health Bureau, so they were classed as
"low risk of bias." The other two studies were journal
Study selection articles25,30 with no declarations of interest or support
Seven electronic databases were searched and 733 arti- and were assessed as "unclear risk of bias."

JTCM | www. journaltcm. com 346 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

CNKI (n = 422)
Wan Fang (n = 170) PubMed (n = 8)
SinoMed (n = 69) CoChrane (n = 16)
VIP (n = 44) EMBASE (n = 4)

Articles identified through database Repeated articles

searching (n = 733) excluded (n = 179)

Studies not eligible for the inclusion criteria

(n = 485):
Articles after duplications removed
Non-RCTs (n = 165);
(n = 554)
Not meeting the diagnostic criteria (n = 227);
Without specific interventions (n = 93)

Studies meeting the exclusion criteria (n = 60):

Not meeting the requirements of decoction
ingredients (n = 46);
Articles for further identification (n = 69) Without complete prescription (n = 5);
No related outcomes (n = 2);
Incorrect data (n = 4);
Incomplete data because of conference
abstracts or protocol (n = 2) and failure to
obtain full text after contacting with the
author (n = 1)
Studies eligible and included (n = 9)

Studies included in quantitative

synthesis (Meta-analysis) (n = 9)

Figure 1 Flow chart of study selection

CNKI: China National Knowledge Infrastructure; VIP: China Science and Technology Journal Database; EMBASE: Excerpta Medica
Database.
Effects of interventions: FSH The remaining study,23 which was not included in the
All of the included studies examined FSH, and a me- Meta-analysis, divided participants into subgroups ac-
ta-analysis could be conducted with data from eight cording to whether FSH levels were normal (< 10 mIU/
studies. Of these, one27 tested TCM combined with mL) or not (10-40 mIU/mL) before treatment and ob-
WM versus WM, and the others24-26,28-31 tested TCM served changes in FSH levels after treatment in different
versus WM. These seven studies showed no heterogene- subgroups. The baselines (FSH levels before treatment)
ity (P = 0.70 > 0.1, I 2 = 0% < 50%), so the fixed-effect of the TCM group and WM group in two subgroups
model was applied. The results showed no obvious dif- were both consistent. In the abnormal FSH subgroup,
ference between TCM and WM in improving FSH lev- FSH level after treatment was（16.74 ± 7.50）mIU/mL
els for patients with DOR (MD 0.11, 95% CI − 0.52 in the TCM group (n = 19) and（14.74 ± 4.59）mIU/
to 0.74, 392 participants, seven trials), and no statisti- mL in the WM group (n = 19). There was no signifi-
cally significant difference (P = 0.73 > 0.05). For the cant difference between the groups (P = 0.32 > 0.05).
study comparing TCM combined with WM versus In the normal FSH subgroup, FSH level after treatment
WM, the results (MD −1.87, 95% CI −3.68 to −0.06, was（8.68 ± 0.71）mIU/mL in the TCM group (n = 11)
60 participants, one trial) showed that the effectiveness and（8.61 ± 0.76）mIU/mL in the WM group (n = 11)
of integrated therapy may be better than only WM and there was no significant difference between the
(P = 0.04 < 0.05). A meta-analysis of these eight stud- groups (P = 0.82 > 0.05).
ies showed no significant difference in FSH levels of Effects of interventions: AMH
the experimental group and control group ( − 0.11, AMH was examined in two studies.24,26 These showed
95% CI − 0.70 to 0.49, 452 participants, eight trials) heterogeneity (P = 0.001 < 0.1, I 2 = 91% > 50%), so
(Figure 3). the random-effect model was used. The results indicat-

JTCM | www. journaltcm. com 347 June 15, 2020 | Volume 40 | Issue 3 |
 Table 1 Characteristics of included studies
Sample size Age
Study Symptom pattern Intervention Control Outcome measure
(T/C) (T/C)
TCMSS, menstruation condition, effective rate
33.77±4.01 Kidney deficiency and Chinese Artificial menstrual cycle (estradiol valerate /
Cai YM 201723 30/30 for TCMSS, pregnancy rate, FSH, LH, FSH/LH,
/34.17±4.21 liver depression medicine estradiol cyproterone)
E2, OV, AFC, AE
Kidney deficiency,
Chinese Artificial menstrual cycle (estradiol valerate / TCMSS, effective rate for TCMSS, pregnancy
Guan Y 201624 30/25 24-39 liver depression and
medicine estradiol cyproterone) rate, FSH, LH, FSH/LH, AMH, OV, AFC, AE
blood deficiency
Combined estrogen and progesterone
Chinese Total effective rate, FSH, FSH at 3 months after

JTCM | www. journaltcm. com

Li SP 200425 35/35 21-40 - (conjugated estrogen +
medicine the end of treatments
gestodene-ethinylestradiol)
29.65±5.12 Kidney deficiency and Chinese Artificial menstrual cycle (estradiol valerate / TCMSS, effective rate for TCMSS, FSH, LH, E2,
Lu YJ 201726 20/20
/30.15±4.67 blood stasis medicine estradiol valerate + progesterone) AMH
Chinese TCMSS, menstruation condition, effective rate
28.23±4.84 Kidney deficiency and Artificial menstrual cycle (estradiol valerate /
Tang S 201627 30/30 medicine+ for TCMSS, FSH, LH, FSH/LH, E2, AFC, OV,
/28.07±4.55 liver depression estradiol valerate + progesterone)
control RI, AE
TCMSS, effective rate for TCMSS, TCMSS at 3
Artificial menstrual cycle (conjugated estrogen / months after the end of treatment, FSH, LH, E2,
Zhang YN 30.23±5.02 Kidney deficiency and Chinese
35/32 conjugated estrogen + medroxyprogesterone FSH at 3 months after the end of treatment, LH
201028 /30.78±4.94 liver depression medicine
acetate) at 3 months after the end of treatment, E2 at 3

348
months after the end of treatment, AE
Artificial menstrual cycle (conjugated estrogen /
Zhong WP 33.35±5.17 Kidney deficiency and Chinese TCMSS, total effective rate, effective rate for
20/20 conjugated estrogen + medroxyprogesterone
201129 /32.40±5.36 liver depression medicine TCMSS, FSH, LH, E2, OVD, AFC, RI, AE
acetate)
Liu LQ et al. / Systematic Review

Zhong WP et al 33.63±4.52 Kidney deficiency and Chinese Artificial menstrual cycle (estradiol valerate / TCMSS, total effective rate, effective rate for
30/30
201730 /33.73±4.92 liver depression medicine estradiol valerate + progesterone) TCMSS, FSH, LH, E2, OVD, AFC, RI, AE
Zhou YC et al 34.47±3.65 Chinese Combined estrogen and progesterone effective rate for TCMSS, KI, FSH, LH, E2, FSH/
30/30 Kidney deficiency
201431 /33.17±4.07 medicine (desogestrel-ethinylestradiol) LH
Notes: AE: adverse events; AFC: antral follicle count; AMH: anti-Müllerian hormone; E2: estradiol; FSH: follicle-stimulating hormone; KI: Kupperman index; LH: luteinizing hormone; OV: ovarian volume;
OVD: ovarian diameter; RI: resistance index; TCMSS: Traditional Chinese Medicine symptom score.
AFC

dex here.
(Figure 4).

(Figure 5).

0.04 < 0.05).

Effects of interventions:

June 15, 2020 | Volume 40 | Issue 3 |

was also significant (P =
perimental group and 5.15
the left ovary AFC was
treatment. After treatment,
was different from ours, so

group, and this difference

± 1.23 in the control
was 5.76 ± 1.02 in the ex-
ovary AFC after treatment
(P < 0.0001). The right
group (n = 30), and this
groups. The criterion used
AFC in the different sub-
WM group (MD 0.55,
higher than that of the
the AFC of the TCM
The results indicated that

difference was significant

3.20 ± 1.54 in the control
mental group (n = 30) and
5.33 ± 1.49 in the experi-
were consistent before
group and WM group
of both sides in TCM
respectively. The baselines
the AFC of each side ovary
Another study27 counted
we did not analyze this in-
and observed changes in
not (< 4) before treatment,
AFC was normal (≥ 4) or
cording to whether the
ipants into subgroups ac-
One study23 divided partic-
participants, three trials)
95% CI 0.05 to 1.04, 155
group was significantly
fect model was applied.
50% ), so the random-ef-
0.03 < 0.1, I 2 = 70% >
showed heterogeneity (P =
ta-analysis. The studies
could be combined for me-
studies and three24,29,30
AFC was examined in five
participants, two trials)
95% CI −0.62 to 1.58, 95
ly different (MD 0.48,
ed that AMH levels of the

group were not significant-

TCM group and the WM
 Liu LQ et al. / Systematic Review

three studies23,29,30 examined TCM versus WM. There

Blinding of participants and personnel (performance bias)

was no heterogeneity in these three studies (P = 0.62 >
0.1, I 2 = 0% < 50%), so the fixed-effect model was ap-

Blinding of outcome assessment (detection bias)

plied. The results indicated that the TCMSS was signif-

Random sequence generation (selection bias)

icantly lower in the TCM group than in the WM

Incomplete outcome data (attrition bias)

Allocation concealment (selection bias)
group (MD −2.39, 95% CI −3.83 to −0.94, 160 partic-
ipants, three trials), indicating that TCM was much

Selective reporting (reporting bias)

more effective than WM in reducing TCMSS for
DOR patients with kidney deficiency and liver depres-
sion pattern (P = 0.001 < 0.01). The results of the
study that assessed the TCMSS in a TCM combined
with WM group versus a WM group showed that the
TCMSS in an integrated therapy group was much low-

Other bias
er (MD − 2.93, 95% CI − 4.89 to − 0.97, 60 partici-
pants, one trial), indicating that integrated therapy was
more effective than WM (P = 0.003 < 0.01). A me-
Cai YM 201723 ta-analysis of these four studies showed that the TC-
MSS was much lower in the experimental group than
Guan Y 201624 in the control group (MD − 2.58, 95% CI − 3.74 to −
Li SP 200425 1.42, 220 participants, four trials), indicating the effec-
tiveness of TCM in this area (P < 0.0001) (Figure 6).
Lu YJ 201726
Effects of interventions: effective rate of TCMSS
Tang S 201627
The effective rate of the TCMSS for the kidney defi-
Zhang YN 201028 ciency and liver depression pattern was examined in
five studies.23,27-30 One study28 was not included in the
Zhong WP 201129
Meta-analysis because of its TCMSS data was incom-
Zhong WP et al 201730 plete. In the rest four studies, one27 examined TCM
combined with WM versus WM, and the other three23,
Zhou YC et al 201431 29,30
examined TCM versus WM. These three studies
Figure 2 Risk of bias summary: review authors' judgments showed no heterogeneity (P = 0.42 > 0.1, I 2 = 0% <
about each risk of bias item for each included study 50% ), so the fixed-effect model was used. The results
showed that the effective rate of TCMSS in the TCM
Effects of interventions: TCMSS group was significantly higher than that in the WM
The TCMSS of the kidney deficiency and liver depres- group (RR 1.18, 95% CI 1.02 to 1.36, 160 partici-
sion pattern was examined in five studies.23,27-30 Of pants, three trials), indicating that TCM was more ef-
these, one study28 reported this incompletely so it was fective than WM for DOR patients with kidney defi-
not included in the Meta-analysis, another study27 ex- ciency and liver depression pattern (P = 0.02 < 0.05).
amined TCM combined with WM versus WM, and The results of the study that examined the effective

Figure 3 Forest plot of follicle-stimulating hormone levels

JTCM | www. journaltcm. com 349 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

Figure 4 Forest plot of anti-Müllerian hormone levels

Figure 5 Forest plot of antral follicle count levels

Figure 6 Forest plot of Traditional Chinese Medicine symptom score

rate of TCMSS in a TCM combined with WM group The results of the study examining TCM combined
and a WM group showed no difference in the effective with WM versus WM showed that the FSH/LH ratio
rate between the two groups (RR 1.13, 95% CI 0.89 to in the integrated therapy group was a little higher than
1.44, 60 participants, one trial, P = 0.32 > 0.05). A me- that in the WM group (MD 0.27, 95% CI 0.01 to
ta-analysis of these four studies showed a higher effec- 0.53, 60 participants, one trial), suggesting a higher ef-
tive rate of TCMSS for the experimental group than fectiveness in the control group (P = 0.04 < 0.05). A
for the control group (RR 1.17, 95% CI 1.03 to 1.32, meta-analysis of these four studies found no signifi-
220 participants, four trials), indicating the superiority cant difference in the FSH/LH ratio between the ex-
of TCM in this area (P = 0.01) (Figure 7). perimental group and the control group (MD 0.1,
Effects of interventions: FSH/LH 95% CI －0.42 to 0.62, 175 participants, three trials)
The FSH/LH ratio was examined in four studies.23,24,27,31 (Figure 8).
As one study23 divided participants into subgroups ac- We did not analyze the index in the remaining study23
cording to whether FSH/LH ratios were normal that was not included in the Meta-analysis, as it used
(≤ 3.6) or not (> 3.6) before treatment and observed 3.6 as the cutoff for differentiating normal or abnor-
ratio changes in different subgroups, the Meta-analysis mal FSH/LH ratios, which was inconsistent with the
was conducted using data from the other three studies. criterion we set.
Of these three, one27 examined TCM combined with
WM versus WM, and the other24,31 examined TCM ver- Effects of interventions: FSH at 3 months
sus WM. These latter two studies showed significant post-treatment
heterogeneity24,31 (P = 0.08 < 0.1, I 2 = 67% > 50%), so Two studies25,28 examined FSH levels at 3 months
we used a random-effect model. When the two studies post-treatment. These showed little heterogeneity (P =
examining TCM versus WM were combined, the re- 0.16 > 0.1, I 2 = 0.48% < 50% ), so the fixed-effect
sults showed no significant difference between the two model was used. The results indicated that FSH at 3
groups in FSH/LH ratio (MD 0.01, 95% CI −0.95 to months post-treatment was significantly lower in the
0.96, 115 participants, two trials, P = 0.99 > 0.05). TCM group than in the WM group (MD −4.77, 95%

JTCM | www. journaltcm. com 350 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

Figure 7 Forest plot of effective rate of Traditional Chinese Medicine symptom score

Figure 8 Forest plot of follicle-stimulating hormone/luteinizing hormone ratios

CI − 6.09 to − 3.45, 137 participants, two trials), indi- Publication bias assessment
cating a superior effect of TCM in this area (P < Funnel plots were constructed to evaluate the bias in
0.000 01) (Figure 9). the included studies (Figure 10). The FSH funnel
graph showed no obvious bias.
Adverse events
Six studies recorded adverse events.23,24,27-30 Most ad- Sensitivity analysis
verse events occurred in the control groups and in- A sensitivity analysis was conducted on FSH, FSH/
volved nausea27,28 (n = 6) and stomachache27,28 (n = LH, AFC, TCMSS, and effective rate of TCMSS,
2). One study23 included brief reports of the follow- which were reported in more than two studies. Studies
ing adverse events (n = 3): dizziness, nausea, loss of were divided into different subgroups according to the
appetite, and distending breast pain, all of which oc- type of literature (journal articles or academic disserta-
curred in the control group. Two studies reported ad- tions) and the combined effect size in different sub-
verse events in the experimental group: nausea27 (n = groups was compared. There were no significant sub-
1) and diarrhea28 (n = 1). The other three studies re- group differences in FSH and TCMSS. There were sub-
ported no adverse events. group differences in combined effect size for FSH/LH

Figure 9 Forest plot of follicle-stimulating hormone levels at 3 months post-treatment

JTCM | www. journaltcm. com 351 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

0
SE (MD) tive in reducing FSH than only WM. TCM was more
effective in improving AFC levels. This may indicate
1
that TCM can fundamentally improve or stimulate
2 ovarian reserve. These outcomes are important indica-
3 tors when diagnosing DOR, so their improvement sug-
4 gests that TCM has some effectiveness for DOR.
MD TCM showed some superiority over WM in improving
5
－4 2 0 2 4 TCMSS and the effective rate of TCMSS. The symp-
TCM vs WM TCM + WM vs WM tom score scales are based on the clinical features of the
Figure 10 Funnel plot of follicle-stimulating hormone levels
kidney deficiency and liver depression pattern. Tonify-
TCM: Traditional Chinese Medicine; WM: Western Medicine;
MD: mean difference. ing-kidney and regulating-liver therapy may be more
effective than WM in relieving clinical symptoms in pa-
ratios, AFC, and the effective rate of TCMSS (Figures tients with kidney deficiency and liver depression,
11-13). which indicates the superiority of syndrome differentia-
tion and treatment. We noticed that TCM also has an
advantage in terms of long-term effectiveness (as in
DISCUSSION FSH tested at 3 months post-treatment). However, per-
We searched seven databases and found nine RCTs ex- haps owing to clinical difficulties, there are not many
amining the effectiveness of kidney-tonifying and liv- studies in this area. We found only two studies that ex-
er-regulating therapy on DOR. The results show that amined this outcome, which affects the validity of the
the effectiveness of kidney-tonifying and liver-regulat- present findings. Future clinical or experimental stud-
ing therapy is not inferior to that of hormone therapy ies are needed to focus on long-term outcomes.
for DOR. The assessment of the adverse effect reporting of the in-
Most of the studies indicated that the effectiveness of cluded studies suggests that TCM is safer than WM.
TCM is comparable to that of WM in improving sexu- The funnel plots showed no obvious bias. The sensitivi-
al hormone levels (FSH, AMH, and FSH/LH). A com- ty analysis indicated instability in the FSH/LH, AFC,
bination of TCM and WM27 seemed much more effec- and effective rate of TCMSS findings, which may re-

Figure 11 Sensitivity analysis of follicle-stimulating hormone/luteinizing hormone ratios

Figure 12 Sensitivity analysis of antral follicle count levels

JTCM | www. journaltcm. com 352 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

Figure 13 Sensitivity analysis of effective rate of Traditional Chinese Medicine symptom score

sult from the difference in literature sources. In addi- REFERENCES

tion, there was a publication bias. The number of stud-
1 Pastore LM, Christianson MS, Stelling J, Kearns WG,
ies included in the Meta-analysis for each outcome
Segars JH. Reproductive ovarian testing and the alphabet
ranged from two to eight (i.e., less than ten). The total
soup of diagnoses: DOR, POI, POF, POR, and FOR. J
sample sizes in the meta-analysis for each outcome
Assist Reprod Genet 2018; 35(1): 17-23.
were insufficient. Additionally, all studies were conduct-
2 Practice Committee of the American Society for Repro-
ed in hospitals of TCM or integrated medicine on the
ductive Medicine. Testing and interpreting measures of
Chinese mainland. All journal articles were written in
ovarian reserve: a committee opinion. Fertil Steril 2015;
Chinese and published in Chinese journals. Therefore, 103(3): e9-e17.
the findings showed some geographical bias, language 3 Levi AJ, Raynault MF, Bergh PA, Drews MR, Miller BT,
bias, and publishing bias. Scott RT Jr. Reproductive outcome in patients with dimin-
None of the included studies used blinding and the al- ished ovarian reserve. Fertil Steril 2001; 76(4): 666-669.
location concealment was unclear. The sample size of 4 Ferraretti AP, La Marca A, Fauser BCJM, Tarlatzis B, Nar-
each group in all the studies was less than 35. The qual- gund G, Gianaroli L, on behalf of the ESHRE working
ity of included studies was generally low. Although ob- group on Poor Ovarian Response Definition. ESHRE con-
jective indicators have been selected to measure effec- sensus on the definition of 'poor response' to ovarian stim-
tiveness, as DOR is manifested as an endocrine disor- ulation for in vitro fertilization: the Bologna criteria. Hum
der that is susceptible to mental factors, blinding is nec- Reprod 2011; 26(7): 1616-1624.
essary for the objectivity of results. Therefore, more 5 Wang CJ, Wu J. The etiology and treatment of dimin-
randomized clinical trials with large sample sizes, multi- ished ovarian reserve. Zhong Guo Shi Yong Fu Ke Yu
center samples, and double-blinding are needed to con- Chan Ke Za Zhi 2015; 31(8): 723-726.
firm that TCM tonifying-kidney and regulating-liver 6 Cohen J, Mounsambote L, Prier P, et al. Outcomes of
therapy is more effective than WM in DOR treatment. first IVF/ICSI in young women with diminished ovarian
In conclusion, the effectiveness of tonifying-kidney reserve. Minerva Ginecol 2017; 69(4): 315-321.
and regulating-liver therapy was not inferior to that of 7 Devine K, Mumford SL, Wu M, DeCherney AH, Hill
hormone therapy in improving the sex hormone MJ, Propst A. Diminished ovarian reserve in the United
States assisted reproductive technology population: diag-
(FSH, AMH, FSH/LH) levels of DOR patients. The
nostic trends among 181 536 cycles from the Society for
effectiveness of tonifying-kidney and regulating-liver
Assisted Reproductive Technology Clinic Outcomes Re-
therapy for increasing AFC, easing accompanying
porting System. Fertil Steril 2015; 104(3): 612-619.
symptoms of patients with a kidney deficiency and 8 Cohen J, Chabbert-Buffet N, Darai E. Diminished ovari-
liver depression pattern, and improving FSH an reserve, premature ovarian failure, poor ovarian re-
long-term is superior to that of hormone therapy. sponder-a plea for universal definitions. J Assist Reprod
However, because of the low quality of the included Genet 2015; 32(12): 1709-1712.
studies, these conclusions constitute insufficient evi- 9 Greene AD, Patounakis G, Segars JH. Genetic associa-
dence for the effectiveness and safety of tonifying-kid- tions with diminished ovarian reserve: a systematic review
ney and regulating-liver therapy in DOR treatment. of the literature. J Assist Reprod Genet 2014; 31(8):
More well-designed RCTs are required to confirm 935-946.
these findings. 10 Shen Y, Ren ML. Effect Factors of Diminished Ovarian

JTCM | www. journaltcm. com 353 June 15, 2020 | Volume 40 | Issue 3 |
 Liu LQ et al. / Systematic Review

Reserve. Shi Yong Fu Chan Ke Za Zhi 2013; 29(9): 22 Liu S, Wu JR, Zhang D, Tan D. What are the best Salvia
643-645. miltiorrhiza injection classes for treatment of unstable an-
11 Han Y, Zhang YS. Logistic regression analysis of related gina pectoris? A systematic review and network Meta-anal-
factors on diminished ovarian reserve. Xian Dai Fu Chan ysis. J Tradit Chin Med 2018; 38(3): 321-338.
Ke Jin Zhan 2015; 24(2): 109-111. 23 Cai YM. Clinical observation on the treatment of Zishens-
12 Gleicher N, Barad DH. Dehydroepiandrosterone (DH-
hugan method upon decreased ovarian reserve patients.
EA) supplementation in diminished ovarian reserve
Kunming: Yunnan TCM University, 2017: 9-24.
(DOR). Reprod Biol Endocrinol 2011; 9: 67.
24 Guan Y. The Clinical observation on the treatment of Bush-
13 Menopause group of Chapter of obstetrics and gynecology
in Chinese Medical Association. Guidelines for the clinical en Shugan Yangxue method upon decreased ovarian reserve
application of menopausal management and hormone re- patients. Beijing: Beijing TCM University, 2016: 40-51.
placement therapy (2012). Zhong Hua Fu Chan Ke Za 25 Li SP. Efficacy of Tonifying-kidney and Smoothing-liver
Zhi 2013; 48(10): 795-799. method in treating 35 patients with diminished ovarian re-
14 Feng JH, Kong LW, Li LL. Research progress of pathoge- serve. Sichuan Zhong Yi 2004; 22(12): 62-63.
ny and treatment of Traditional Chinese and Western 26 Lu YJ. Analysis of the related influencing factors of AMH
Medicine of decreasing ovarian reservation. Zhong Guo Yi and the intervention effect of Bushen Huoxue Decoction.
Yao Dao Bao 2014; 11(14): 161-164. Nanjing: Nanjing TCM University, 2017: 24-29.
15 Xia T, Ma RH, Mu W, et al. Traditional Chinese Medi- 27 Tang S. The Clinical Research of Yishentiaochong Tang
cine for diminished ovarian reserve: a systematic review combined artificial cycle in treating kidney deficiency and
and Meta-analysis. Chinese Herbal Medicines 2014; 6(2):
liver depression menstruation late result from diminished
93-102.
ovarian reserve. Jinan: Shandong TCM University, 2016:
16 Ma WW, Xu LW. Distribution of Chinese medical syn-
2-10.
dromes of diminished ovarian reserve. Zhong Yi Yao Lin
Chuang Za Zhi 2018; 30(6): 1068-1071. 28 Zhang YN. The Clinical observation of Bu Shen Shu Gan
17 Zhang XF, Huang HT, Li YB. Distribution rules of TCM method for improving decreasing ovarian reserve. Harbin:
Syndromes of decline in ovarian reserve in Beijing based Heilongjiang TCM University, 2010: 26-33.
on potential category analysis. Beijing Zhong Yi Yao 2017; 29 Zhong WP. Clinical study of Bu-shen-huo-xue decoction
36(9): 789-792. on treating decreased ovarian reserve on account of psycho-
18 Chen ZJ, Tian QJ, Qiao J, et al. Chinese expert consensus logical stress. Nanjing: Nanjing TCM University, 2011:
about the diagnosis and treatment of premature ovarian in- 14-24.
sufficiency. Zhong Hua Fu Chan Ke Za Zhi 2017; 52(9): 30 Zhong WP, Wang PJ, Ye YQ, Wang J, Ni YY. Clinical ob-
577. servation of Kidney-supplimenting and blood-quickening
19 China Pharmacopoeia Committee. Pharmacopoeia of the
decoction on the treatment of ovarian dysfunction caused
People's Republic of China (2015). Beijing: China Medi-
by psychological factors. Hebei Zhong Yi 2017; 39(2):
cal Science Press, 2015: 1-422.
203-207.
20 Gao XM. Traditional Chinese Pharmacy. 2nd ed. Beijing:
China Press of Traditional Chinese Medicine, 2007: 1-547. 31 Zhou YC, Yin XQ. Clinical Effect of "Bushen Tiaojing de-
21 Higgins JPT, Altman DG, Gøtzsche PC, et al. The Co- coction" in treating 30 patients with diminished ovarian
chrane Collaboration's tool for assessing risk of bias in ran- reserve. Shanghai Zhong Yi Yao Da Xue Xue Bao 2014; 28
domised trials. BMJ 2011; 343: d5928. (5): 38-41.

JTCM | www. journaltcm. com 354 June 15, 2020 | Volume 40 | Issue 3 |