The basal PCOS panel — FSH, LH, oestradiol and total testosterone — is drawn on day 2 to 4 of the cycle, counting the first day of bleeding as day 1. Progesterone is the exception and goes mid-luteal, about 7 days before the next period is due. 17-hydroxyprogesterone must be before 8am. Prolactin and TSH have no cycle day at all, and AMH is broadly cycle-independent.
Written by Dr. Chaithra H. A, Consultant · Obstetrics & Gynaecology (Minimal access surgery), AJSMC, Egmore, Chennai (TNMC 102973). This page is about how PCOS is tested for and what the results can and cannot say. Everything below follows the 2023 International Evidence-based Guideline for the Assessment and Management of Polycystic Ovary Syndrome (Teede HJ et al., 2023, peer reviewed across 39 partner and collaborating organisations, including ESHRE, ASRM and the Endocrine Society), which replaced the 2018 update and carries 254 recommendations and practice points.
One thing before the tests, because the symptoms that bring women to a PCOS page are not all PCOS. Very heavy or prolonged bleeding, bleeding that causes dizziness or fainting, and sudden severe pelvic pain are not a hormone-panel problem. They need a hospital with a 24-hour emergency department, or a call to 108. AJSMC is an outpatient centre in Chennai with no emergency department, no inpatient beds and no intensive care, and it is closed outside Monday to Saturday, 10am to 9pm, and all day Sunday. Testing for PCOS is never urgent. Bleeding heavily at 2am is.
Which blood test goes on which day of your cycle?
This is the table the rest of the page rests on. Day 1 is the first day of proper bleeding, not spotting. If you are not bleeding at all, none of the follicular-phase timings can be applied by you at home. That is a conversation with the gynaecologist, who will either wait for a bleed or work around it.
| Test | What it is for | When in the cycle | Time of day | Fasting required by guideline? |
|---|---|---|---|---|
| FSH | Excluding hypogonadotrophic hypogonadism, usually from low body fat or intensive exercise — not to diagnose PCOS | Early follicular phase, day 2–4 | Morning | No |
| LH | Read alongside FSH for the same exclusion | Early follicular phase, day 2–4 | Morning | No |
| Oestradiol | Interpreting the FSH result, not a test in its own right | Early follicular phase, day 2–4, drawn with FSH | Morning | No |
| Progesterone | Confirming whether ovulation actually happened | Mid-luteal — day 21 of a 28-day cycle, or about 7 days before the period is expected; later for longer cycles | Any | No |
| Total and free testosterone | Biochemical hyperandrogenism — one of the three Rotterdam criteria | Early follicular phase, with the basal panel | Morning | No |
| 17-hydroxyprogesterone | Excluding nonclassic congenital adrenal hyperplasia | Early follicular phase | Early morning, before 8am | No |
| Prolactin | Excluding hyperprolactinaemia | No cycle-day requirement | Any | No |
| TSH | Excluding thyroid disease, which mimics the cycle disturbance of PCOS | No cycle-day requirement | Any | No |
| AMH | May stand in for the ultrasound in adults, to define polycystic ovarian morphology | Broadly cycle-independent, though levels may vary across the cycle | Any | No |
Sources for the timings, test by test: the ASRM Practice Committee opinion on ovarian reserve testing (2020) for the day 2 to 4 basal sampling of FSH, LH and oestradiol; NICE fertility guidance for the mid-luteal progesterone; the Endocrine Society congenital adrenal hyperplasia guideline (2018) for the early-morning 17-hydroxyprogesterone; and the 2023 International Evidence-based PCOS Guideline for AMH. Testosterone is booked into the same early-follicular morning slot as the rest of the basal panel, and the requesting doctor will specify the exact window on the form. The fasting column is short because these are hormone tests, not metabolic ones, the tests that do involve preparation come later, after a diagnosis, and the consultant who orders them will say what preparation each one needs.
Why is a test drawn on the wrong day wasted?
Because for several of these tests, the number moves so much across the cycle that a sample taken at the wrong moment produces a result that is real, printed, normal-looking and meaningless. It does not come back flagged as unusable. It comes back looking like an answer, and it sends the next decision in the wrong direction.
| Test | What goes wrong on the wrong day or hour | Source |
|---|---|---|
| FSH | FSH varies both within a cycle and between cycles, which limits how much a single measurement can be trusted; a sample taken mid-cycle catches the surge instead of the baseline | ASRM Practice Committee opinion, 2020 |
| LH | The mid-cycle LH surge makes any non-basal reading uninterpretable | ASRM Practice Committee opinion, 2020 |
| Oestradiol | An early rise in oestradiol pulls an otherwise elevated FSH back into the normal range, so the FSH result is misread as normal. An oestradiol above 60–80 pg/mL with a normal FSH may itself indicate ovarian dysfunction | ASRM Practice Committee opinion, 2020 |
| Progesterone | Drawn too early it is low in everyone, so it proves nothing about ovulation | NICE fertility guidance |
| 17-hydroxyprogesterone | A random or late-day sample can be normal even in nonclassic congenital adrenal hyperplasia, which is exactly the condition it was ordered to exclude | Endocrine Society, 2018 |
The progesterone row deserves its own paragraph, because it is the single commonest wasted test in this whole workup and PCOS is the reason. The classic "day 21 progesterone" is calibrated to a 28-day cycle. PCOS is defined partly by cycles longer than 35 days or fewer than 8 a year, so in precisely the population being tested, day 21 lands in the follicular phase, weeks before ovulation could have happened, and returns a low value that says nothing at all. NICE's own instruction is to take the sample 7 days before the expected period — day 21 of a 28-day cycle, day 28 of a 35-day cycle, and to interpret it once the next period has begun. A woman in Chennai who books a day-21 progesterone off a generic checklist, on a 45-day cycle, has bought a number that cannot be interpreted.
One more practical trap, from recommendation 1.2.6 of the 2023 guideline. If you are already taking a combined oral contraceptive pill, tell the doctor before any androgen test is booked. The pill raises sex hormone-binding globulin and reduces gonadotrophin-dependent androgen production, which makes biochemical hyperandrogenism very difficult to assess reliably. The guideline says that where the androgen assessment is imperative, the pill has to be withdrawn for a minimum of three months with contraception managed otherwise in that time, a decision for the consultant to make with you, never one to take on your own before an appointment.
What has to be excluded before PCOS can be diagnosed?
PCOS is never diagnosed on its features alone. Every version of the criteria, from Rotterdam 2003 to the 2023 guideline, requires that other causes be excluded first, and the guideline's own diagnostic algorithm names the panel exactly. This is why the blood test list is longer than the two or three hormones most people expect.
| What must be excluded | Test used | Timing |
|---|---|---|
| Thyroid disease | TSH | Any time |
| Hyperprolactinaemia | Serum prolactin | Any time |
| Nonclassic congenital adrenal hyperplasia | 17-hydroxyprogesterone | Early morning, before 8am, early follicular phase |
| Hypogonadotrophic hypogonadism, usually from low body fat or intensive exercise | FSH and LH, with clinical assessment | Early follicular phase |
| Cushing's syndrome, adrenal or ovarian androgen-secreting tumours | Only where clinically indicated — particularly with rapid onset, or androgens markedly above the laboratory reference range | As directed |
Source: the exclusion footnote to Algorithm 1 of the 2023 International Evidence-based PCOS Guideline. Note what is not on that list. Fasting insulin and HOMA-IR are sold routinely as part of PCOS packages across India, and recommendation 1.9.12 of the 2023 guideline is explicit that clinically available insulin assays "are of limited clinical relevance and are not recommended in routine care". An LH:FSH ratio of 2:1 or 3:1 is not on the list either. It is not a diagnostic criterion in the 2003 Rotterdam consensus, the 2018 update or the 2023 guideline, and LH and FSH appear in the current algorithm only to exclude something else.
On the laboratory itself: the guideline asks for validated LC-MS/MS assays for total testosterone rather than direct immunoassays, which it says "have limited accuracy and demonstrate poor sensitivity and precision for diagnosing hyperandrogenism in PCOS", and it asks for free testosterone to be assessed by calculation, equilibrium dialysis or ammonium sulfate precipitation rather than measured directly. If an androgen result is borderline and the treatment decision hangs on it, the method used is a fair question to ask the requesting doctor.
Do I need a scan at all?
Often not, and this is the most useful practical line in the 2023 guideline for anyone in Chennai about to book one. Recommendation 1.4.9 states that in patients with irregular menstrual cycles and hyperandrogenism, an ovarian ultrasound is not necessary for a PCOS diagnosis. Recommendation 1.5.2 says the same about AMH. If both of those features are already present and other causes have been excluded, the two criteria needed are met and a scan adds nothing to the diagnosis.
| Rotterdam criterion, as updated in 2023 | What counts |
|---|---|
| Ovulatory dysfunction | Cycles shorter than 21 or longer than 35 days, or fewer than 8 cycles a year, from 3 years after menarche to perimenopause; or any single cycle longer than 90 days |
| Hyperandrogenism | Hirsutism at a modified Ferriman-Gallwey score of 4–6 depending on ethnicity, or a raised total or calculated free testosterone |
| Polycystic ovarian morphology | 20 or more follicles in at least one ovary on a transvaginal scan; or an ovarian volume of 10 ml or more; or, in adults only, serum AMH used in place of the scan |
Two of the three are required, other causes must be excluded, and ultrasound and AMH should not both be done — recommendation 1.5.5 says so explicitly, to limit overdiagnosis. Note the follicle threshold, because it is the number most commonly quoted wrongly: 20 or more follicles per ovary is the current adult threshold, and the older figure of 12 comes from the 2003 consensus and is superseded. Where older equipment or poor image quality makes counting follicles across the whole ovary unreliable, an ovarian volume of 10 ml or more, or 10 or more follicles per section, is used instead, and a transabdominal scan is reported primarily on volume for the same reason. Where a scan is indicated and acceptable to the woman, the transvaginal route is the most accurate.
And the word "polycystic" is a historical misnomer worth correcting. What the radiologist counts are follicles — normal structures that every ovary contains, not cysts. A scan line reading "polycystic ovaries" describes one of three possible criteria. On its own it is not a diagnosis, and plenty of women with that line on a report do not have PCOS.
When should an AMH test be done, and can it diagnose PCOS?
It can be done on any day, and no, it cannot diagnose PCOS. That is the whole answer, and it is worth stating flatly because AMH is marketed harder than any other test in this workup. Recommendation 1.5.1 allows serum AMH for defining polycystic ovarian morphology in adults; recommendation 1.5.3 states that AMH "should not be used as a single test for the diagnosis of PCOS"; and recommendation 1.5.4 says it should not yet be used in adolescents at all.
AMH's freedom from cycle timing is relative rather than absolute. Practice point 1.5.6 of the same guideline lists what moves it: age, with levels peaking between 20 and 25 years; BMI, with lower levels at higher BMI; current or recent combined oral contraceptive use, which may suppress it; and the menstrual cycle itself, across which AMH may vary. Laboratories are asked to use population-specific and assay-specific cut-offs, which is why no single AMH number can be printed here as a threshold, the same value from two different assays does not mean the same thing.
What counts as an irregular cycle, exactly?
Many women decide their cycles are irregular, or decide they are not, on the basis of how the last two months felt. The guideline puts numbers on it, and they change with how long ago periods started.
| Years since the first period | A cycle is irregular if it is |
|---|---|
| Under 1 year | Not applicable — irregularity in the first year is a normal part of the pubertal transition |
| 1 to under 3 years | Shorter than 21 days or longer than 45 days |
| 3 years to perimenopause | Shorter than 21 days or longer than 35 days, or fewer than 8 cycles in a year |
| Over 1 year, at any age | Any single cycle longer than 90 days |
| No period at all | By age 15, or more than 3 years after breast development started |
Source: recommendation 1.1.1 of the 2023 guideline. Two consequences follow. First, an adolescent within a year of her first period is not a candidate for this workup, and one who has features but does not meet the criteria is considered at increased risk and reassessed at or before full reproductive maturity, 8 years after menarche. Neither ultrasound nor AMH is used in that age group. Second, a regular cycle does not by itself rule out anovulation: recommendation 1.1.5 says ovulatory dysfunction can still occur with regular cycles, and where anovulation needs confirming, that is what the mid-luteal progesterone is for.
What else is checked once PCOS is diagnosed?
The metabolic screen, and it starts at diagnosis rather than at some later stage. The 2023 guideline is unusually firm about this, and Indian data explains why: among 1,224 Indian women with PCOS in an ICMR-funded nationwide study published in 2024, 91.9% had dyslipidaemia, 43.2% had obesity by Asian criteria, 32.9% had non-alcoholic fatty liver disease, 24.9% had metabolic syndrome, 8.3% had hypertension and 3.3% had diabetes.
| Check | When first | How often after |
|---|---|---|
| Lipid profile — total cholesterol, LDL, HDL, triglycerides | At diagnosis, regardless of age and BMI | Based on whether hyperlipidaemia and other risk factors are present |
| Blood pressure | At diagnosis | Annually, and when planning pregnancy or fertility treatment |
| Glycaemic status, ideally a 75-g oral glucose tolerance test | At diagnosis, in adults and adolescents, regardless of BMI | Every 1 to 3 years, based on individual risk factors |
| Fasting glucose and/or HbA1c | Only where an oral glucose tolerance test cannot be done | Same interval, but the guideline notes significantly reduced accuracy |
| Fasting insulin or HOMA-IR | Not recommended in routine care | — |
Source: recommendations 1.8.3, 1.8.4, 1.9.2, 1.9.3, 1.9.9, 1.9.10 and 1.9.12 of the 2023 guideline. An oral glucose tolerance test is also considered in all women with PCOS without pre-existing diabetes when planning pregnancy or fertility treatment, and offered at 24–28 weeks in pregnancy. Obstetrics & Gynaecology and Diabetology are both published departments at AJSMC in Egmore, Chennai, the reproductive side of PCOS and the metabolic side are not separate illnesses.
How common is PCOS in India?
It depends entirely on which criteria are used, and any page that gives you one number without naming the criteria has skipped the most important part of the answer.
| Prevalence | Figure | Criteria used |
|---|---|---|
| India, weighted national | 19.6% (95% CI 12.7–29.2) | Rotterdam 2003 |
| India, weighted national | 7.2% (95% CI 4.8–10.8) | NIH 1990 |
| Global range, for comparison | 10–13% | Rotterdam |
The Indian figures come from the same ICMR-funded cross-sectional study of 8,993 women aged 18–40 across five geographical zones, recruited between November 2018 and July 2022 and published in 2024. Same women, same data, nearly threefold difference — driven purely by the definition applied. The global range is from recommendation 1.6.1 of the 2023 guideline, which also notes that prevalence is broadly similar across world regions but may be higher in South East Asian and Eastern Mediterranean populations.
What does treatment actually start with?
Lifestyle, and the guideline is careful about how it says so. Recommendation 3.1.1 recommends lifestyle intervention — exercise alone, or a multicomponent approach combining diet, exercise and behavioural strategies — for all women with PCOS, to improve metabolic health including central adiposity and lipid profile. Recommendation 3.1.3 calls lifestyle management a core focus of PCOS management. Recommendation 3.1.5 is the one that matters most for anyone who has tried and stalled: there are benefits to a healthy lifestyle even in the absence of weight loss.
That sentence is in the guideline for a reason, and it is a safety point rather than a kindness. Recommendation 2.5.1 requires that eating disorders and disordered eating be considered in PCOS regardless of weight, particularly in the context of weight management and lifestyle interventions, and the guideline also asks that anxiety and depression be assessed. A page that reduces PCOS to a weight target is unsafe for a meaningful share of the women reading it. Goals are meant to be co-developed with the woman and to respect her own preferences, and behavioural work — goal-setting, self-monitoring, problem solving, relapse prevention — is part of the recommendation rather than an optional extra.
Beyond lifestyle, what is offered depends on which problem is actually troubling you, the cycle, the skin and hair, the metabolic risk, or fertility, and those are different conversations with different answers. Nothing about that choice can be made from a webpage, and no medicine is named on this one. It is the consultant's decision, made with you, after the tests above have been read together.
When should you see a doctor?
Book an outpatient appointment if your cycles fit the irregular definitions in the table above, if unwanted hair growth or persistent acne is bothering you, if you have been trying to conceive without success, or if you are simply holding a scan report with the words "polycystic ovaries" on it and want to know whether it means anything. Bring the dates of your last two or three periods, because the first thing that has to be decided is which day of your cycle each test can be drawn on. Bring any previous hormone reports with their dates, and tell the doctor about any hormonal contraception you are taking.
Ask to be seen sooner, and say so when you call, if the androgen symptoms came on quickly or are worsening fast — new or rapidly progressive hair growth, a deepening voice, or other changes suggesting virilisation. Recommendation 1.2.9 and practice point 1.3.7 of the 2023 guideline treat rapid onset as the clinical clue that separates PCOS from an androgen-secreting tumour of the ovary or adrenal gland, from Cushing's syndrome and from congenital adrenal hyperplasia. That picture needs assessment, not a routine hormone panel booked for whenever is convenient.
Do not come to AJSMC, or wait for it to open, if you are bleeding heavily — soaking through a pad an hour, or feeling faint with the bleeding, or in sudden severe pelvic pain. Go to a hospital with a 24-hour emergency department, or call 108. AJSMC is outpatient only, Monday to Saturday, 10am to 9pm, with no emergency department and no beds, and this is one of the questions where knowing where not to go matters as much as knowing where to go. For a PCOS assessment itself, the Obstetrics & Gynaecology outpatient at AJSMC on Police Commissioner Office Road, Egmore, Chennai 600008 can be reached on 044 2532 2021.
