LH & FSH: What studies support—and what remains unclear

LH & FSH: What studies support—and what remains unclear

LH (luteinizing hormone) and FSH (follicle-stimulating hormone) are the central gonadotropins that help drive oocyte or sperm maturation via the hypothalamus–pituitary–gonads axis. Studies mainly provide clues as a consequence of other interventions (e.g., in PCOS), rather than treating LH/FSH “optimization” itself as the direct target in high-quality long-term RCTs. That’s exactly why interpretation is so challenging.

What LH and FSH do in the body (and why “optimization” is hard)

LH and FSH are signaling molecules that help shape maturation in the ovaries and testes through endocrine control loops. The key point: many studies don’t measure LH/FSH to test a specific “target value,” but because LH/FSH are downstream markers that change under therapies. This does not automatically mean that every desired LH/FSH shift is beneficial.

In practice, you often see the same reasoning: “If LH/FSH shows X, the problem must be Y—and if we specifically change LH/FSH, Y improves.” Biologically, this is usually not that straightforward. First, LH/FSH are dynamic: pulse frequencies, time-of-day patterns, and feedback mechanisms (including via sex steroids and possibly inhibin) can strongly affect measured levels. Second, LH/FSH are tightly integrated into the body’s broader “architecture”: if you change insulin resistance, inflammation, body weight, or energy availability, multiple signals typically shift at once—and LH/FSH are only one piece.

So the most important methodological point for the evidence is this: There are few RCTs that optimize LH/FSH as a direct target with long-term endpoints (e.g., better fertility, fewer miscarriages, fewer tumor events). More often, researchers assess whether an intervention (e.g., Metformin or an acupuncture-based therapy for PCOS) improves reproductive endocrine parameters, including gonadotropin-adjacent measures. That is scientifically reasonable—but it does not automatically answer the question, “Which LH/FSH values are ideal?”

If you try to derive concrete “target ranges” from that, you always need the matching study logic: Which population? Which intervention? Which endpoints? This specific mapping is what differentiates the evidence later when looking at Metformin/acupuncture and contraception.

Lifestyle before supplements: What studies more often reflect in the hormonal axis

If you want to understand LH/FSH as part of an indirect system, the best starting point is lifestyle: sleep, movement, morning light, and metabolic/weight management can influence the hormonal environment more strongly than individual supplements. For PCOS, systematic reviews also suggest that dietary interventions can change metabolic and endocrine endpoints—though the robustness for fine LH/FSH goals may be limited depending on the question.

Why lifestyle first? Not because supplements are inherently “bad,” but because the hormonal axis is highly sensitive to higher-level regulators: insulin/glucose, fat distribution, stress and circadian mechanisms, as well as inflammatory and energy signals. Especially in PCOS, the metabolic component is central—so it’s more likely that LH/FSH patterns will shift under metabolically effective interventions without LH/FSH itself being “dosed.”

From the nutrition side, it matters what the systematic review on ketogenic diets in overweight/obesity and PCOS concludes: (Diha et al., 2026, PMID 41853422). The paper addresses metabolic, endocrine, and reproductive outcomes. Still, a sober interpretation is essential: such a review does not automatically mean that every direction of LH or FSH change is consistently shown and clinically meaningful in all studies. Systematic reviews can be heterogeneous (different dietary implementations, different PCOS definitions, different study durations).

If you neglect lifestyle “basics,” additional interventions aimed at the hormonal axis are often a worse first step, because it becomes harder to separate what truly drives change. If you’re interested in sleep rhythm as a methodological comparison, see: Circadian rhythm: effects & evidence (what is supported). And if you’re thinking about macronutrients/timing, the evidence filter provided by Carbohydrate periodization: effects & evidence up to meta-analysis can help too—even if it isn’t specifically about LH/FSH.

In practice: before thinking about “LH/FSH support,” build a structural foundation first (metabolic status, weight status, cycle stability). Supplements—if considered at all—should be a secondary add-on, never the primary lever.

Evidence hierarchy: meta-analysis, RCT, observational study—and what it means for LH/FSH

Meta-analyses and RCTs usually provide the most reliable statements about effects under interventions. For LH/FSH, that means: if you want higher evidence, pay particular attention to meta-analyses of PCOS treatments, because gonadotropin changes are systematically summarized there as outcomes. Observational studies are more useful for association questions (links), not for causal “therapy recommendations” aimed at LH/FSH target values.

At a high level, the evidence hierarchy is roughly: RCTs minimize confounding, meta-analyses increase precision, and observational data show patterns but do not prove causality. For LH/FSH, this is especially relevant because gonadotropins often sit between cause and effect: they are part of the endocrine axis that itself is influenced by metabolism, sex steroids, and possibly therapeutic interventions.

For PCOS, you can see a fairly clear picture: meta-analyses bundle interventions such as Metformin or acupuncture-related therapies across trials that often report reproductive endocrine outcomes—and therefore also gonadotropin-adjacent measures. The strength comes from aggregation: it’s not a single study that decides, but the overall pattern.

• For Metformin, there is an explicit large systematic review and meta-analysis available on gonadotropins: (Krysiak et al., 2026, PMID 41891336). This is methodologically relevant because it includes 51 randomized controlled trials.
• For acupuncture, systematic reviews or network meta-analyses exist that compare therapies and address, among other things, endocrine and morphological parameters: (Du Z et al., 2026, PMID 41837127).

Single studies are still useful—but mainly for mechanistic plausibility or for identifying which patient profiles might respond. For “target ranges” or “optimization,” evidence from single studies alone is usually too weak.

Observational studies help when the question is risk associations. Example: in the STEED study (Wu et al., 2026, PMID 41485106), the focus is on sex steroid hormones and gonadotropins in relation to testicular germ cell tumors. This can suggest patterns, but it’s not a guarantee that changing LH/FSH later causally reduces or increases risk.

In short: Meta-analysis/RCT = “What happens under an intervention?” Observation = “What is associated with what?” You need to apply this separation consistently when discussing LH/FSH, otherwise “marker changes” quickly turn into unsupported therapy promises.

Study overview: Which questions the evidence actually answers

What the best PCOS studies show: Metformin & acupuncture and their effects on gonadotropins

For PCOS, the strongest evidence is when you treat and then measure gonadotropin changes as an outcome—rather than treating LH/FSH as the direct “target.” In meta-analyses, Metformin and acupuncture-related therapies have repeatedly shown effects on endocrine parameters; but that does not mean you can derive a precise “LH/FSH target profile” from them without further assumptions.

Metformin: Gonadotropin changes as a bundled RCT outcome

The meta-analysis by Krysiak et al. includes 51 randomized controlled trials and reports the effect of Metformin on gonadotropin levels in women with PCOS (Krysiak et al., 2026, PMID 41891336). The strength is the bundling: you tend to see a more consistent direction than in individual small RCTs. However, there is an important limitation: meta-analyses typically report average effects on endocrine measurements, but they don’t automatically provide clinical translation for every individual (e.g., fertility, miscarriage risk, or long-term safety in each subgroup).

Without quoting the exact numerical results block of this specific meta-analysis here, you can still state methodologically: the evidentiary weight comes from the number of RCTs and the systematic review design—not from a single “trial signal.” For your decision logic: if your question is, “How often and in what direction does Metformin change gonadotropin-adjacent parameters in PCOS?”, this is one of the strongest sources available in your list.

Acupuncture: Network meta-analysis instead of a single signal

For acupuncture, the evidence picture is similar, but interpretation depends even more on which acupuncture protocols and which endpoints were measured in the RCTs. The network meta-analysis by Du Z et al. (Du Z et al., 2026, PMID 41837127) summarizes acupuncture-related therapies and evaluates, among other things, reproductive endocrine outcomes and ovarian morphology. The paper is valuable because it doesn’t only assess “acupuncture vs control,” but compares therapies within a network.

Again, the key caveat applies: this supports endocrine effects under specific interventions, but it does not replace direct RCT evidence showing that you should optimize LH/FSH “directionally” for your goals. It may instead mean that acupuncture indirectly affects processes that then shift gonadotropin patterns.

If you make concrete intervention decisions (e.g., Metformin or complementary therapies), the ideal basis is a clinical indication (insulin resistance, cycle problems, fertility goals)—not LH/FSH values as the sole controlling variable.

Contraception and the HPG axis: differential effects on gonadotropin-adjacency and amenorrhea

Contraceptives can influence the hypothalamus–pituitary–gonads axis differently and change the incidence of amenorrhea. This matters because gonadotropin changes under contraceptives can be measurable—yet “controlled” does not automatically mean “health-promoting” or aimed at LH/FSH optimization.

The study by Bick et al. examines injectable contraceptives and reports that they differentially affect the HPG axis and the incidence of amenorrhea (Bick et al., 2026, PMID 41489365). The practical value of this work is less about “How can I optimize LH/FSH up/down?” and more about “What happens biologically under a specific contraceptive method?”

Why this is relevant: many people interpret hormone lab values during contraception as a kind of “function check” of the axis. But contraception is not a physiological therapy for normalization in the sense of restoring a natural regulatory pattern—it is a pharmacological steering of cycle behavior with its own logic (suppression/modification of cycle processes). That gonadotropins (or very closely related axis markers) can decrease or pulse differently is biomedically plausible, but it must be read in the clinical context.

Also, amenorrhea is not simply a “lab artifact.” If the incidence differs by contraceptive type (as described in Bick et al.), then the risk/benefit trade-off depends on the method. For you, that means: if you switch due to symptoms/bleeding patterns or measure labs, you need clear mapping—which method, which population, and which objective (bleeding control, cycle regulation, PCOS symptoms, contraception).

That is also why studies on contraception are more appropriate as reference for “What happens under method X?” rather than “Which supplements increase LH/FSH meaningfully?” If you later look for interactions between substances and endocrine axes, this evidence filter can help: Interactions: what studies show (and what they don’t).

Bottom line: contraception provides valid information about axis responses and side-effect profiles (e.g., amenorrhea), but it is not a direct blueprint for LH/FSH target values in everyday life or in PCOS without the relevant clinical indication.

LH/FSH beyond PCOS: cycles in adolescence and hormonal patterns in cancer risk

Outside PCOS, studies suggest that LH/FSH-adjacent hormonal patterns can vary in the context of developmental stages and risk-associated endpoints. The data mainly help with understanding associations—not as a basis to derive therapeutic LH/FSH target values.

Adolescence and irregular cycles: PCO vs multifollicular

Villa et al. investigated insulin- and GH-adjacent endpoints in a clinical study in adolescent girls with irregular cycles, comparing polycystic vs multifollicular ovarian trajectories (Villa et al., 2003, PMID 12841537). The methodological value: adolescent cycle regulation differs from adults—it is developmentally variable. Accordingly, hormonal patterns in this age group are harder to interpret as a simple “error state.”

For LH/FSH questions, that means: if you see lab values in adolescence, the question of “optimization” is often methodologically misframed, because normality itself depends on definitions (maturation stage, trajectory, dynamics). Villa et al. provides more context here on endocrine-axis environments and how they connect to metabolic/regulatory signals.

STEED study: gonadotropins and the risk of testicular germ cell tumors

Wu et al. report from the STEED study about sex steroid hormones, gonadotropins, and their relationship with the risk of testicular germ cell tumors (Wu et al., 2026, PMID 41485106). The crucial interpretation is this: these study designs primarily generate associations, not causality. Even if gonadotropins correlate with risk groups in certain comparisons, that does not mean that a targeted change in LH/FSH would be preventive. For that, you would need causal intervention data—which is not available in this evidence form.

For informed lay decision-making, this matters because the topic is sometimes misused in “hormone optimization” forums: “If we change the marker, we change the risk.” That is not cleanly supported here. What the study more accurately does: generate hypotheses and direct attention to hormonal development and risk factors.

So if you consider LH/FSH in a risk context, use the data as information about patterns—not as a direct action plan. The distinction between “association” and “treatment effect” should be your guardrail.

What you should take away from this

• LH/FSH are markers in an axis, not automatically a controllable regulator with clear target ranges; many studies measure them as a result of other interventions.
• For PCOS, the evidence is strongest when you look at treatments such as Metformin (Krysiak et al., 2026, PMID 41891336) or acupuncture-related therapies (Du Z et al., 2026, PMID 41837127) in an RCT/meta-analysis context—not as an isolated LH/FSH target-value strategy.
• Lifestyle and metabolic levers often matter more in practice, because they shift the hormonal environment broadly; ketogenic diet is an example: a systematic review exists, but the robustness for fine LH/FSH targets is not always equal (Diha et al., 2026, PMID 41853422).
• Contraception shows differential effects on the axis and amenorrhea (Bick et al., 2026, PMID 41489365)—useful for “what happens under method X?” rather than automatically informing “how do I optimize LH/FSH?”.