Therapeutic effects of chromium supplementation on women with polycystic ovarian syndrome: A systematic review and meta-analysis

Abstract

Background

Polycystic ovarian syndrome (PCOS) has been treated recently with chromium supplementations. However, it is unknown if this dietary supplement has similar effect to metformin.

Aim

The aim of this study was to investigate the efficacy of chromium supplementation in women with PCOS.

Methods

A meta-analysis was conducted using relevant articles obtained from searches of PubMed, Scopus, Embase, and Google Scholar. The mean difference and standardized mean difference were employed to determine the effect size for biochemical parameters.

Results

A total of 10 randomized controlled trials involving 683 women were included in the analysis. The results indicated that chromium supplementation, as vs a placebo, significantly decreased fasting blood insulin (P = 0.01), triglyceride (P < 0.00001), total cholesterol (P < 0.00001), very low-density lipoprotein (P < 0.00001), low-density lipoprotein (P = 0.0003), high sensitivity C-reactive protein (P = 0.02), malondialdehyde (P = 0.007), follicle stimulating hormone (P = 0.0007), and prolactin (P = 0.01), and increased the Quantitative Insulin Sensitivity Check Index (P = 0.02), total antioxidant capacity (P < 0.0001), and ovulation incidence (P = 0.001). Chromium supplementation was also found to be more effective than metformin in reducing HOMA-IR (P < 0.00001), and luteinizing hormone (P = 0.04).

Conclusion

Chromium picolinate supplementation at a dosage of 200 μg may provide benefits similar to metformin with regard to FBG, FBI, ovulation, and pregnancy incidence, with fewer side effects in patients with PCOS. Further experiments are still required to draw effective dietary guidelines related to chromium.

Resumen

Antecedentes

El síndrome de ovario poliquístico (SOP) se ha tratado recientemente con suplementos de cromo. Sin embargo, se desconoce si este suplemento dietético tiene un efecto similar al de la metformina.

Objetivo

El objetivo de este estudio fue investigar la eficacia de los suplementos de cromo en mujeres con SOP.

Métodos

Se realizó un metaanálisis utilizando artículos relevantes obtenidos de búsquedas en PubMed, Scopus, Embase y Google Scholar. Se emplearon la diferencia de medias y la diferencia de medias estandarizada para determinar el tamaño del efecto de los parámetros bioquímicos.

Resultados

Se incluyeron en el análisis un total de 10 ensayos controlados aleatorizados con 683 mujeres. Los resultados indicaron que los suplementos de cromo, en comparación con un placebo, disminuyeron significativamente la insulina en sangre en ayunas (p = 0,01), los triglicéridos (p < 0,00001), el colesterol total (p < 0,00001), las lipoproteínas de muy baja densidad (p < 0,00001), las lipoproteínas de baja densidad (p = 0,0003), la proteína C reactiva de alta sensibilidad (p = 0,02), el malondialdehído (p = 0,007), la hormona foliculoestimulante (p = 0,0007) y la prolactina (p = 0,01), y aumentaron el índice cuantitativo de comprobación de la sensibilidad a la insulina (p = 0,02), la capacidad antioxidante total (p < 0,0001) y la incidencia de ovulación (p = 0,001). Los suplementos de cromo también resultaron más eficaces que la metformina para reducir el HOMA-IR (p < 0,00001) y la hormona luteinizante (p = 0,04).

Conclusión

La suplementación con picolinato de cromo a una dosis de 200 μg puede proporcionar beneficios similares a la metformina en lo que respecta a la FBG, la FBI, la ovulación y la incidencia de embarazo, con menos efectos secundarios en pacientes con SOP. Aún es necesario realizar más experimentos para elaborar directrices dietéticas eficaces en relación con el cromo.

Introduction

Polycystic ovary syndrome (PCOS) is one of the most common endocrine disorders in women of reproductive age, affecting up to 20% of women, depending on the diagnostic criteria and ethnicity. PCOS is characterized by a metabolic disorder in which disturbances in sex hormones and insulin resistance occur.^1,2 Approximately 75% of women with PCOS are insulin-resistant.³ PCOS is associated with several complications including lower rates of pregnancy, infertility, irregular cycles, hyperandrogenism, and hirsutism. Moreover, women with PCOS are at a higher risk of developing metabolic syndrome, type 2 diabetes mellitus, cardiovascular disease, and cancer.^4,5

The main causative factors of PCOS development remain unclear. However, some events are thought to interact with one another, leading to PCOS development. Elevated levels of luteinizing hormone (LH) and gonadotropin-releasing hormone (GnRH) occur, along with decreased levels of follicle-stimulating hormone (FSH). In healthy people, the main function of these hormones is to regulate the levels of sex hormones such as testosterone and estrogen in the bloodstream. LH is responsible for testosterone synthesis in ovarian thecal cells, whereas FSH is responsible for stimulating the secretion of estrogen and the conversion of excess androgens to estrogens by producing aromatases. This disturbance in the LH/FSH ratio results in hyperandrogenism, which is very common in patients with PCOS.⁶ In addition, LH is thought to act synergistically with insulin.⁷ Mechanistic studies have found that excess insulin secretion activates LH receptors, resulting in excess testosterone synthesis.^8,9 In addition, these high levels of LH were shown to increase serum free testosterone (FT) levels by decreasing the synthesis of sex hormone-binding globulin (SHBG) and finally amplify hyperandrogenism.¹⁰

Moreover, hyperandrogenic women with PCOS have higher insulin resistance than women with lower levels of hyperandrogenic PCOS.¹¹ Excess insulin secretion might also result in excess fat storage, leading to overweight or obesity, which worsens the condition. Regardless of PCOS-induced insulin resistance, up to 88% of women with POCS were shown to be either overweight or obese, which is another factor causing insulin resistance.¹² Some studies have found that women with PCOS and obesity have higher levels of fasting insulin and blood glucose levels vs weight-matched control participants.^13,14 Excess body fat mainly contributes to PCOS development through insulin resistance, causing inflammatory markers, such as IL-6 and TNF-α.¹⁵ Moreover, achieving a 5% reduction in body weight in these patients has been shown to increase insulin sensitivity and improve blood glucose concentrations.^16–18

Therefore, the first-line therapy for PCOS is to improve diet quality by following specific dietary approaches such as a Mediterranean diet or a low-glycemic index diet, aiming to reduce body weight and subsequently insulin resistance. Moreover, lipid profile, hormonal profile, and inflammatory markers have shown to improve with dietary interventions and weight loss in patients with PCOS.^19,20 Furthermore, most PCOS patients exhibit poor food behaviors and low adherence to specific dietary approaches.^21,22 For this reason, recent studies have focused on compensating for micronutrient deficiencies that increase insulin sensitivity and improve glucose metabolism in PCOS patients. For instance, women with PCOS have lower zinc levels than healthy women.^23,24 Zinc supplementation increases insulin sensitivity and reduces HOMA-IR in these patients.²⁵

Similarly, chromium (Cr) is a micronutrient that is deficient in women with PCOS according to a recent study.²⁶ It is thought that Cr can increase insulin sensitivity, leading to blood glucose and insulin homeostasis. In patients with diabetes, Cr is efficient in improving the metabolic profile, including blood glucose and insulin levels.²⁷ A previous meta-analysis reported contradictory results when assessing the effect of Cr on fasting blood insulin (FBI) levels in women with PCOS.^28,29 This is mainly due to the small sample size and the small number of studies reported. Therefore, the aim of this study was to assess the effect of Cr supplementation on the metabolic profile, hormonal profile, and inflammation status of patients with PCOS.

Materials and methods

We conducted a meta-analysis based on the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).³⁰ The study protocol was registered with PROSPERO (CRD42024513304).

Search strategy

A search of online databases including PubMed, Embase, and Scopus was conducted up to the cutoff date of 04/01/2024. We conducted a systematic search across the available scientific literature to identify clinical trials that assessed the effect of Cr supplementation in PCOS patients on metabolic profile, hormonal profile, and inflammation status. We used these keywords in our search (polycystic ovary OR polycystic ovary syndrome OR pcos OR polycystic ovarian syndrome OR polycystic ovary disease OR polycystic ovarian disease OR stein leventhal syndrome OR sclerocystic ovarian degeneration OR ovarian degeneration OR ovaries OR insulin sensitizing OR insulin resistan* OR type 2 diabet* OR t2d) AND (chromium) AND (women OR female OR woman).

Study selection

We reviewed the titles and abstracts of non-duplicated articles. We used a screen form with a hierarchical approach based on the study design, population, exposure, and outcome to exclude articles that did not meet the inclusion criteria. To identify any additional studies, we reviewed the reference lists of the included studies. In addition, a manual search was performed. The full texts of eligible citations were then reviewed.

Inclusion and exclusion criteria

Inclusion criteria:

• (A)

  Chromium intervention in the oral form.

• (B)

  Cr plus another supplement.

• (C)

  Randomized clinical trials.

• (D)

  Peer-reviewed articles.

• (E)

  Rotterdam diagnostic criteria.

• (F)

  Patients with PCOS and other comorbidities (chronic diseases).

• (G)

  Studies with placebo or metformin as control group.

• (H)

  Patients aged 18–55 years

• (I)

  Articles in English language.

Exclusion criteria:

• (A)

  Intervention or control includes a combination of >1 drug.

• (B)

  Patients without PCOS

• (C)

  and not English articles.

• (D)

  Not peer-reviewed articles.

• (E)

  Patients <18 years and >55 years

• (F)

  Studies with <15 patients

Data extraction

After the article selection process, with regard to the inclusion and exclusion criteria, the following information (first author's last name, publication date, study location, total sample size, age, diagnostic criteria, study design, intervention dose, control, study duration, and main outcomes) was extracted from the articles and listed in Table 1.

Table 1.

Characteristics of included study.

First author's last name, year, country	Sample size I/C	Age (mean)	Diagnostic criteria	Study design	Intervention and dose	Control	Duration	Main results
Jamilian, 2015, Iran	60 (30/30)	N/R	Rotterdam	DB-RCT	200 μg Cr	Placebo	2 months	ProlactinFSH↓LHDHEAFT17-OH-progesteroneMF-G score↓Hs-CRP↓NOTAC↑GSHMDA↓
Jamilian and Asemi, 2015, Iran	64 (32/32)	25	Rotterdam	DB-RCT	200 μg Cr	Placebo	2 months	BMIFBGFBI↓HOMA-IR↓HOMA-B↓QUICKI↑TG (P = 0.05)TC (P = 0.05)VLDLLDLHDL
Ashoush, 2015, Egypt	85 (44/41)	25	Rotterdam	DB-RCT	1000 μg Cr	placebo	6 months	FBGFBI↓FGIR↑TTF-G scoreBMI↓OvulationPregnancy
Siavashani, 2018, Iran	40 (20/20)	34	Rotterdam	DB-RCT	200 μg Cr	Placebo	2 months	BMIHs-CRP↓NOPregnancy rate
Jamilian, 2018, Iran	40 (20/20)	32	Rotterdam	DB-RCT	200 μg Cr	Placebo	2 months	BMIFBG↓FBI↓HOMA-IR↓QUICKI↑TG↓TC↓VLDL↓LDLHDLTAC↑GSHMDA↓MF-G score
Jamilian, 2019B, Iran	54 (27/27)	28.5	Rotterdam	DB-RCT	200 μg Cr + 1000 mg carnitine	Placebo	3 months	BMI↓FBG↓FBI↓HOMA-IR↓HOMA-BQUICKI↑TG↓TC↓VLDL↓LDL↓HDLTotal-/HDL cholesterol ratio↓
Jamilian, 2019, Iran	54 (27/27)	N/R	Rotterdam	DB-RCT	200 μg Cr + 1000 mg carnitine	Placebo	3 months	TT↓SHBG (P = 0.07)MF-G score↓Hs-CRP↓NOTAC↑GSHMDA↓
Amooee, 2013, Iran	92 (46,46)	27	Rotterdam	DB-RCT	200 μg Cr	Metformin 500 mg × 3 a day	3 months	BMIFBG↓FBI↓QUICKI↑TTFTFSH (P = 0.07)LHLH/FSHProlactin (P = 0.06)TSH (P = 0.07)Ovulation ratePregnancy rateSide effects ↓
Borg, 2022, Egypt	140 (70,70)	28	Rotterdam	DB-RCT	1000 μg Cr	Metformin 500 mg × 3 a day	3 months	BMIFBG↓MFBI↓MFTProgesterone (P = 0.06)Ovulation ratePregnancy rateSide effects↓
Kishk, 2019, Egypt	54 (27,27)	28	Rotterdam	RCT	200 μg Cr	Metformin 500 mg × 2 a day	3 months	BMI↓Hirsutism↓FBG↓FBI↓HOMA-IR↓QUICKI↑FT↓FSH↓LHProlactinTSHOvulation ratePregnancy rate

Arrows refer to increase or decrease. Arrows with M refer to metformin group.

I: intervention; P: placebo; DB-RCT: randomized clinical trial double blinded.

Quality assessment

The quality of the included studies was evaluated using the Cochrane Collaboration tool. This tool includes the following key parts: random sequence generation, allocation concealment, blinding, incomplete outcome data, and selective reporting. Each item was categorized as having low, unclear, or high risk of bias. Accordingly, studies with more than two items of low risk were categorized as studies with good quality, studies with two items of low risk were considered studies with fair quality, and those with <2 items were considered studies with low risk of bias.³¹

Statistical analysis

The present meta-analysis was performed using the Cochrane Program Review Manager Version 5.4. Variables assessed in 3 or 4 studies were included. In this regard, net changes in the mean ± SD of body mass index (BMI), fasting blood glucose (FBG), FBI, Homeostasis Model Assessment for Insulin Resistance (HOMA-IR), Quantitative Insulin Sensitivity Check Index (QUICKI), triglyceride (TG), total cholesterol (TC), very low-density lipoprotein (VLDL), low-density lipoprotein (LDL), high-density lipoprotein (HDL), (FT), total testosterone (TT), prolactin, FSH, LH, high sensitivity C-reactive protein (Hs-CRP), nitric oxide (NO), total anti-oxidant capacity (TAC), malondialdehyde (MDA), glutathione (GSH), and Ferriman–Gallwey (F-G) score, and events of ovulation, pregnancy, and side effects were assessed. In studies where net changes were not directly reported in the intervention and control groups, the final measurement values were collected. We preferred to use net changes rather than final measurements to increase the precision. In case all studies within a forest plot reported values with the same unit (mean difference), we combined studies that used changes and final measurements. However, if data were reported in different units and we had to use standardized mean difference, we used either the final measurement or change in the same forest plot as recommended by the Cochrane online handbook. In any case reporting the standard error of the mean (SEM), standard deviation (SD) was calculated using the following formula: SD = SEM × sqrt (n), being n the number of participants. The random effects model was applied for pooling analysis to compensate for the heterogeneity of the studies.^31,32 Interstudy heterogeneity was explored quantitatively using Cochran's Q and I² statistics. I² ≤50% and ≥75% indicated substantial and considerable heterogeneity, respectively.³³P-values were considered statistically significant at <0.05. Sources of heterogeneity were explored using sensitivity analysis. We conducted two sensitivity analysis to detect any impact of dose (studies >200 μg were excluded) and duration (studies >2 months were excluded) on the heterogeneity.

Results

Literature selection

Overall, 805 citations were obtained from the initial search (Fig. 1). All randomized controlled trials that explored the effects of Cr on patients with PCOS were included. A total of 678 of articles remained after excluding duplicates. A total of 655 of articles were excluded based on their titles or abstracts. A total of 23 of articles were eligible for inclusion in this systematic review and meta-analysis. Thirteen out of the 23 studies of interest were excluded for various reasons (explained in the flow chart). The remaining 10 studies were included in the qualitative and quantitative analyses.

Figure 1.

PRISMA 2020 flow diagram illustrating the systematic review and meta-analysis process.

Risk of bias assessment

Risk of bias assessment was performed in all the included studies. Assessment revealed that all studies were of a low risk of bias (high quality). The detailed results of each item are shown in (Supplementary Fig. 1). Risk of bias summary is shown in (Supplementary Fig. 2).

Characteristics of the studies

A total of 10 RCTs were included in this meta-analysis.^34–43 The 10 RCTs included 683 women with a mean age of 28. A total of 7 studies were conducted in Iran^34,36–41 and 3 in Egypt.^35,42,43 All study participants were women diagnosed with PCOS based on the Rotterdam criteria. The formula of Cr in all studies was chromium picolinate. The intervention included a total of supplementation at doses of 200 μg,^{34,36–39,42} 1000 μg,³⁵ and 200 μg chromium supplementation + 1000 mg carnitine.^40,41 Control groups were either placebo,^35–41 or metformin.^34,42,43 Dose of metformin was 1500 mg daily^34,43 and 1000 mg daily.⁴² The intervention went on between 2 and 6 months. All studies were randomized, double-blinded controlled trials, except⁴² which was not blinded. One author published five different papers.^36,40,41 The BMI was reported in 8 studies.^{34,35,37–39,41–43} Fasting blood glucose and fasting blood insulin were reported in 8 studies.^{34,35,37,38,41–43} Triglyceride, total cholesterol, VLDL, LDL, and HDL were reported in 3 studies.^37,38,41 Free testosterone was reported in 4 studies.^34,36,42,43 FSH, LH, and prolactin levels were reported in 3 studies.^34,36,42 Total testosterone was reported in 3 studies.^34,35,40 Hs-CRP and NO were reported in 3 studies.^36,39,40 TAC, GSH, and MDA levels were reported in 3 studies as well.^36,38,40

Effect of Cr supplementation on BMI

Eight studies examined the effect of Cr supplementation on BMI vs placebo and metformin. Five studies assessed the effect of Cr supplementation on BMI vs placebo, with a total of 283 participants. There were no significant inter-group differences (MD = −0.30, 95% CI, −0.64, 0.03, P = 0.08, I² = 92%). Three studies compared the effects of Cr supplementation and metformin on BMI. The study included a total of 286 participants in both groups. No significant reduction in BMI in the metformin group vs the Cr group was ever reported (MD = 0.46, 95% CI, −0.14, 1.05, P = 0.13, I² = 74%) (Fig. 2).

Figure 2.

Effect of Cr supplementation on BMI.

Effect of Cr supplementation on FBG and FBI levels

Four studies assessed the effect of Cr supplementation vs placebo on FBG and FBI. This study included 243 participants. There was no significant difference in FBG (MD = −1.45, 95% CI, −3.96, 1.06, P = 0.026, I² = 65%) and a significant reduction in FBI (MD = −2.30, 95% CI, −3.47, −1.12, P = 0.0001, I² = 69%) levels in the Cr-supplemented group vs the placebo group. Three studies assessed the effect of Cr supplementation vs metformin on FBG and FBI. This study included 286 participants. There were no significant differences in FBG (MD = −2.92, 95% CI, −17.65, 11.81, P = 0.7, I² = 98%) and FBI (MD = −0.91, 95% CI, −3.87, 2.04, P = 0.55, I² = 95%) levels in the Cr-supplemented group vs the metformin group. When all data were pooled together, Cr supplementation had no effect on FBG reduction (MD = −2.05, 95% CI, −7.01, 2.92, P = 0.42). In contrast, there was a significant reduction in FBI levels (MD = −1.85, 95% CI, −3.24, −0.45, P = 0.010) (Fig. 3).

Figure 3.

Effect of Cr supplementation on FBG and FBI levels.

Effect of Cr supplementation on QUICKI and HOMA-IR

Five studies assessed the effects of Cr supplementation on HOMA-IR and QUICKI scores. Three studies with a total of 158 participants compared Cr with placebo. There was no significant difference in HOMA-IR (MD = −0.29, 95% CI, −0.87, 0.30, P = 0.34, I² = 95%); however, a significant increase in QUICKI was observed (MD = 0.01, 95% CI, 0.00, 0.02, P = 0.02, I² = 84%). One study assessed the effect of Cr supplementation vs metformin on HOMA-IR, and 2 studies assessed QUICKI. There was a significant reduction in HOMA-IR (MD = −0.53, 95% CI, −0.69, −0.37, P < 0.00001, I² = 0%), and no difference was found in QUICKI (MD = 0.02, 95% CI, −0.01, 0.04, P = 0.26, I² = 99%). When all data were pooled together, Cr supplementation had no effect on HOMA-IR (MD = −0.35, 95% CI, −0.78, 0.09, P = 0.12), but a significant increase in QUICKI was observed (MD = 0.01, 95% CI, 0.00, 0.02, P = 0.002) (Supplementary Fig. 3).

Effect of Cr supplementation on TG, TC, VLDL, LDL, and HDL

Three studies examined the effect of Cr supplementation vs placebo on lipid profiles. This study included 158 participants. Cr supplementation caused significant reduction in TG (MD = −24.91, 95% CI, −33.73, −16.10, P < 0.00001, I² = 0%), TC (MD = −15.78, 95% CI, −21.90, −9.66, P < 0.00001, I² = 21%), VLDL (MD = −4.98, 95% CI, −6.74, −3.22, P < 0.00001, I² = 0%), LDL (MD = −9.35, 95% CI, −14.30, −4.31, P = 0003, I² = 14%), and no effect on HDL was ever reported (MD = −0.45, 95% CI, −3.12, 2.23, P = 0.74, I² = 0.84) vs placebo group (Supplementary Fig. 4).

Effect of Cr supplementation on inflammation status

A total of 4 studies were included in the studies that examined the effect of Cr supplementation on inflammation status. Three studies assessed each of the biomarkers. The total number of participants for each biomarker is 159. Cr supplementation caused a significant reduction in Hs-CRP (MD = −1.19, 95% CI, −2.18, −0.20, P = 0.02, I² = 65%), MDA (MD = −0.48, 95% CI, −0.83, −0.13, P = 0.007, I² = 79%), and a significant increase in TAC (MD = 179.74, 95% CI, 98.56, 260.92, P < 0.0001, I² = 61%) vs the placebo group. No effect of Cr supplementation on NO (MD = −0.71, 95% CI, −4.12, 2.69, P = 0.66, I² = 70%), and GSH was ever reported (MD = 19.07, 95% CI, −16.22, 54.35, P = 0.29, I² = 14%) vs placebo group (Supplementary Fig. 5).

Effect of Cr supplementation on FT and TT

Four studies assessed the effect of Cr supplementation on serum FT levels, with a total of 173 participants in each group. When all data were pooled together, there was a significant reduction in FT levels in the control group (SMD = 0.57, 95% CI, 0.08, 1.05, P = 0.02). Three studies assessed the effect of Cr supplementation on serum TT levels, with a total of 117 in the intervention group and 114 in the control group. When all data were pooled, no differences were reported between the Cr and control groups (MD = −0.02, 95% CI, −0.11, 0.07, P = 0.66) (Supplementary Fig. 6).

Effect of Cr supplementation on FSH and LH levels

Three studies assessed the effects of Cr supplementation on serum LH and FSH levels. Overall, 146 participants were included in this study. One study found no effect of Cr supplementation on LH levels vs placebo (MD = 2.80, 95% CI, −3.72, 9.32, P = 0.40). In contrast, compared with metformin, there was a significant reduction in LH levels in the Cr-supplemented group (MD = −0.32, 95% CI, −0.63, −0.01, P = 0.04, I² = 0%). The same overall effect was observed when all data were pooled together. Cr supplementation significantly reduced FSH levels vs placebo (MD = −2.60, 95% CI, −4.11, −1.09, P = 0.0007). No difference was found when Cr supplementation was compared with metformin (MD = −0.76, 95% CI, −2.90, 1.38, P = 0.49, I² = 98%) or when all data were pooled together (MD = −1.29, 95% CI, −3.13, 0.54, P = 0.17) (Supplementary Fig. 7).

Effect of Cr supplementation on prolactin and F-G scores

Three studies assessed the effects of Cr supplementation on prolactin levels. When all data were pooled together, no differences were reported between the Cr and control groups (SMD = 0.02, 95% CI, −0.69, 0.74, P = 0.95). Five studies assessed the effect of Cr supplementation on hirsutism based on the F-G score. When all data were pooled, no differences were reported between the intervention and the control group (SMD = 0.01, 95% CI, −0.23, 0.25, P = 0.93) (Supplementary Fig. 8).

Effect of Cr supplementation on ovulation and pregnancy rates

Five studies assessed the effect of Cr supplementation on ovulation and pregnancy. Two studies compared Cr supplementation with placebo. Cr supplementation was efficient in increasing ovulation number (OR, 3.57, 95% CI, 1.65, 7.76, P = 0.001, I² = 69%). There was no effect when Cr supplementation was compared with metformin (OR, 0.95, 95% CI, 0.61, 1.48, P = 0.82, I² = 0%). No significant difference was found when all data were pooled together (OR, 1.33, 95% CI, 0.92, 1.94, P = 0.13). Cr supplementation was not effective in increasing the pregnancy rate vs placebo (OR, 2.37, 95% CI, 0.71, 7.86, P = 0.16, I² = 12%) or metformin (OR, 0.82, 95% CI, 0.47, 1.43, P = 0.48, I² = 0%) (Supplementary Fig. 9).

Effect of Cr supplementation vs metformin on side effects

Two studies compared the effects of Cr supplementation and metformin on side effects in 232 participants. Cr supplementation showed a significantly higher rate of loss of appetite and a significantly lower rate of abdominal discomfort, diarrhea, vomiting, and nausea vs metformin. The rate of headache was higher in the metformin group vs the Cr group. When all data were pooled together, Cr supplementation was more effective in reducing side effects vs metformin (Supplementary Fig. 10).

Sensitivity analysis

Our sensitivity analysis was performed based on dose and duration where possible as indicated in (Supplementary Table 1). High heterogeneity levels disappeared when considering the dose and duration for BMI, HOMA-IR, NO, and TAC. However, results remain insignificant except for TAC.

Discussion

Summary and explanation

Polycystic ovary syndrome (PCOS) is a complex metabolic disorder characterized by hormonal disturbances, infertility, insulin resistance, and elevated serum lipid levels. Women with PCOS are at a higher risk of developing type 2 diabetes mellitus and cardiovascular diseases due to longitudinally elevated insulin levels. Obesity amplifies the risk by further increasing insulin resistance and blood lipids, such as TG and LDL, which are independent factors that increase the risk of cardiovascular disease. This systematic review and meta-analysis aimed to synthesize evidence on the effects of Cr supplementation in patients with PCOS. Our main results indicated that Cr supplementation could improve the metabolic profile, lipid profile, hormonal profile, and inflammation status. Cr supplementation proved effective in reducing FBI, HOMA-B, TG, TC, VLDL, LDL, hs-CRP, MDA, FSH, and FT. It was also effective in increasing QUICKI, TAC, and ovulation vs placebo. Furthermore, Cr supplementation was more effective in reducing HOMA-IR and LH than metformin and was as effective as metformin in reducing FBG, FBI, FT, TT, FSH, and F-G scores, and increasing QUICKI, prolactin, ovulation, and pregnancy rates. In addition, Cr has fewer side effects than metformin.

Excess body weight is one of the main factors that cause PCOS. In our study, we observed that Cr supplementation was not effective in reducing BMI. However, two of the included studies reported a significant decrease in BMI in the intervention vs the placebo group.^35,41 These reductions are mostly mediated by co-factors. For instance, unlike the other studies, Ashoush advised participants to eliminate their sugar, saturated fats, and total caloric intake, and increase their physical activity, which was repeated at each visit. In the Jamilian study, caloric intake was equal in both groups according to patient records, but Cr was mixed with l-carnitine, which might have influenced body weight reduction. Despite Cr caused higher events of loss of appetite vs metformin, there was higher reduction in BMI in metformin group.^34,43

Moreover, Cr supplementation showed no difference in FBG levels vs placebo and metformin. Despite the significant reduction in BMI in Ashoush (2016) study, no changes were reported in FBG levels. In contrast, Jamilian (2019B) found a reduction in BMI, followed by FBG. When Cr was vs metformin with respect to FBG levels, no differences were found. Studies that reported approximately similar effect sizes reported contradictory results.^42,43 In the Borg (2022) study, metformin was found to be more efficient in reducing FBG levels. However, in the Kishk (2019), Cr proved more effective vs metformin. A similar trend was observed for the FBI. Among the studies that compared Cr with metformin, Kishk (2019) used the lowest dose of metformin (500 mg twice a day), which might explain the results. Furthermore, all studies that compared Cr supplementation with placebo resulted in a significant reduction in FBI, regardless of FBG levels. This result was confirmed with QUICKI, in which all studies showed increased insulin sensitivity in the Cr vs the placebo group. Therefore, our results indicate that Cr supplementation may improve FBI by increasing insulin sensitivity in patients with PCOS.

Unsurprisingly,⁴² showed significant improvement in the intervention group vs metformin regarding QUICKI. Borg (2022) did not run the test, but most probably contradictory results would have been found if the test would have been performed, as was shown for FBG and FBI. In addition, PCOS patients have demonstrated to have elevated serum lipids, such as LDL, which is an independent risk factor for cardiovascular diseases.⁴⁴ Cr supplementation was effective in reducing TG, TC, VLDL, and LDL vs placebo, whereas no effect was observed on HDL. This effect was independent of weight loss. In addition, several studies have reported that PCOS patients exhibit high levels of oxidative stress and inflammation.⁴⁵ In our systematic review and meta-analysis, we observed a significant improvement in inflammation status by the reduction of Hs-CRP and MDA, and an increase in TAC, indicating an anti-inflammatory effect of Cr.

Hormonal profile PCOS is mainly characterized by hormonal disturbances that lead to hyperandrogenism, thereby affecting fertility and hirsutism. Some studies suggest that LH/FSH values in women with PCOS reach 2/1 and even 3/1.⁴⁶ However, baseline values of LH/FSH participants in the included studies did not reach 2/1, suggesting normal LH/FSH serum levels. Exceptionally, the baseline ratio exceeded 2/1 in the intervention group, and both FSH and LH levels were above normal range.³⁶ Furthermore, our results showed that Cr supplementation was effective in reducing LH levels vs metformin. This effect is possibly mediated by a reduction in FBI, as discussed previously. Jamilian (2019) reported a considerable increase in SHBG levels, which may also be caused by a reduction in FBI. In addition, Cr supplementation was effective in reducing FSH vs placebo and was not effective vs metformin.⁴⁰

Moreover, the reduction in FT in the Cr group was significant when data were pooled together, and no difference in prolactin levels was found. However, the results of FT and prolactin might not be representative of the actual effects for several reasons. First, the studies reported their data in different units, which gave highly variable values when converted. Second, we used the final measurements, which in some cases did not present true effects. For example, in a study by Jamilian (2015), the baseline value in the placebo group was significantly lower than that in the baseline value of intervention group. At the end of the trial, no changes were observed in either group. Therefore, comparing the final values between groups yielded unrealistic results. In addition, Cr supplementation had no effect on TT.

Hirsutism is defined as the abnormal distribution of terminal hair in women. Hirsutism is the primary reflection of hyperandrogenism.⁴⁷ It was measured using the F-G score, which is a scoring system ranging from 0 (no excessive terminal hair growth visible) up to 4 (extensive hair growth visible) for 11 different body parts. Modified F-G score reduced the body parts down to 9. Therefore, a maximum score of 36 or 44 is possible, but a score ≥8 is classified as hirsutism.⁴⁸ We found that Cr supplementation had no effect on F-G score. In a study by Jamilian (2015), a significant reduction in the F-G score was observed, despite no reduction in androgens. Jamilian (2019) showed a significant reduction in the F-G score along with TT with carnitine and Cr co-supplementation. This beneficial effect may be related to increased β-oxidation in oocytes.⁴⁹

Anovulation and infertility are other reflections of hyperandrogenism.⁵⁰ We found that Cr supplements, but not metformin, significantly increased the ovulation rate vs placebo. These results may be explained by the study duration. Ashoush (2016) observed a change in ovulation rates in the 4th month and increased gradually until the 6th month. On the other hand, all included studies, except for the one conducted by Ashoush (2016), did not exceed 3 months. In addition, the improvement in ovulation was followed by slightly more pregnancy events in the intervention group. However, we could not include this in our meta-analysis due to insufficient data.

Of note, heterogeneity was high in most results. We performed sensitivity analysis to understand the source of heterogeneity. In case of BMI, HOMA-IR, NO, and TAC the heterogeneity was significantly reduced when we adjusted the studies according to dose (200 μg) and duration (2 months). Although a couple of sensitivity analysis were applied, heterogeneity was not diminished in all variables and the source of heterogeneity is unknown. This could be due to the low number of available studies, conducting more RCTs could clarify the sources of heterogeneity. Furthermore, we aimed to perform further sensitivity analysis according to population and baseline characteristics. However, these choices were inapplicable.

Comparison with other results

Our results are consistent with those of other studies on FBI and BMI.^51,52 In contrast, there was no change in TT, FT, DHEA, FSH, and LH levels after 4 months of supplementation.⁵¹ A dose of 1000 μg Cr supplementation for 6 months in 14–17 aged girls did not impact BMI or hirsutism. However, a significant reduction in FT was observed.⁵³ Similar to our findings, a previous meta-analysis found a reduction in FBI,²⁸ and a recent network meta-analysis indicated that Cr supplementation improved FBI and FBG. In addition, they found no effect of Cr supplementation on HDL and TT levels.⁵⁴

Strengths and Limitations

This systematic review and meta-analysis provide comprehensive findings on the effects of Cr supplementation in PCOS patients. Our study had several strengths, including comprehensive research and appropriate statistical analyses. Additionally, our results are not limited to placebo but also to metformin. However, this study has several limitations. First, the number of included trials was insufficient to draw practical conclusions. Second, all the studies were conducted in Iran and Egypt. Third, the duration of the included studies was relatively short. Fourth, most studies did not measure sex hormones, which are the core of the disease. Lack of sexual hormones measurement could be attributed to research financial issues. In addition, the values of sex hormones in the included studies were extremely variable, which might be due to differences in methods. Dietary intake and physical activity were not consistently accounted for in the included trials. Variability in the dietary pattern of included participants such as caloric intake and macronutrients composition might influence the observed effect of Cr supplementation. For example, caloric restriction or adherence to specific dietary pattern such Mediterranean diet or low-glycemic index diet could independently improve insulin sensitivity and inflammatory markers. Furthermore, the geographic and culture homogeneity of the included trials (Egypt and Iran) where both regions share similar dietary pattern could affect the metabolic outcomes in PCOS patients. Traditional diets of these regions consist of high carbohydrate intake from bread and rice, moderate protein intake, and relatively low-fat intake which ultimately affect glucose metabolism and insulin sensitivity. In addition, micronutrients intake including zinc and Cr may differ due to dietary recourses specific to these regions leading to an impact chromium baseline levels and the observed effects of supplementation. Moreover, cultural attitude towards physical activity may also vary between the regions and other populations. For example, the lower levels of physical activity of certain Middle Eastern populations could amplify insulin resistance which render chromium supplementation more pronounced in these cohorts. These regional dietary and lifestyle patterns limit the generalizability of the findings to populations with different cultures or dietary context.

Future research and recommendations

Future studies should focus on conducting double-blind clinical trials comparing Cr supplementation at different doses for a longer duration. In addition, more comprehensive studies comparing Cr and metformin are required. It might be suggested to combine Cr with metformin in severe PCOS cases, which might provide additional benefits, as shown by Fogle et al.⁵² Future research should concentrate on the effect of Cr supplementation on sex hormones, especially FT, FSH, LH, and prolactin.

Conclusions

The current meta-analysis elaborates on the potential benefits of Cr supplementation in improving metabolic, hormonal and inflammatory parameters in women with PCOS. However, the promising outcomes of Cr supplementation remain limited. Only a small number of studies have directly compared Cr supplementation with metformin and these studies are varied in design, dosage, and duration. Therefore, claiming equivalence or superiority of Cr over metformin is still premature and require studies to better evaluate the comparative safe and efficacy profile of these interventions.

CRediT authorship contribution statement

M.H. conducted the concept, design, and drafting of this study. Y.R. and A.F. searched databases, and screened articles. M.H. was involved in data acquisition, extraction, analysis, and interpretation. Y.R. and A.F. critically revised the manuscript. All authors approved the manuscript final version. M.H. are the guarantors of this study.

Funding

None declared.

Conflicts of interest

None declared.

Acknowledgements

The authors would like to extend their sincere appreciation to the libraries and facilities at Istanbul Yeni Yüzyıl University, United Arab Emirates University, and Al-Ain University. The extensive collection of academic resources, both in print and digital formats, and the assistance of the library staff in accessing and navigating relevant databases greatly enhanced the depth and quality of the current study.