Manage genital tuberculosis and the HIV-gynaecology interface

In one line

Genital tuberculosis is a paucibacillary, scarring disease of the upper genital tract — tubes first, then endometrium — that in South Africa presents as infertility rather than infection, is confirmed by endometrial sampling and a high-specificity but low-sensitivity GeneXpert, and is cured microbiologically by standard six-month therapy while leaving fertility largely unrecoverable; the HIV-gynaecology interface is the second half of the same epidemic, where immunosuppression accelerates cervical neoplasia, magnifies common infections, and turns every gynaecological encounter into an HIV care encounter governed by drug interactions and the rules of HIV gynaecology at registrar level.

Mechanism & pathophysiology

Mycobacterium tuberculosis almost never reaches the genital tract directly. The organism seeds the upper tract haematogenously or via lymphatics from a distant primary focus, usually a pulmonary or pleural one that may have healed and become radiologically silent years earlier; direct extension from intra-abdominal or peritoneal TB is the secondary route, and sexual acquisition from infected semen is described but rare. The blood-borne route explains the anatomical hierarchy that every consultant must hold, because it dictates where to sample and what to expect on imaging: the fallopian tubes are involved in about 90% of cases, the endometrium in roughly 70%, and the ovaries in about 25% — the tubes are the richly vascularised first landfall, the endometrium is seeded next, and the ovary, with its tougher capsule, is reached last.

The disease that follows is one of fibrosis, not suppuration. Tubal mucosa is destroyed and the tube fibroses into a rigid, beaded structure with cornual obstruction and, where the fimbrial end seals, a hydrosalpinx or pyosalpinx; the classic radiological signatures — the lead-pipe or beaded tube, the tobacco-pouch and golf-club deformities, mural calcification — are the imaging correlate of caseating granulomas healing by scar. In the endometrium, repeated granulomatous shedding and basal-layer destruction produce intrauterine synechiae and a shrunken, deformed cavity — an Asherman-like end-state reached by infection rather than instrumentation, which is why amenorrhoea or hypomenorrhoea is a presenting feature and why even a successful embryo transfer later fails to implant. The endometrial granulomas are also why the disease is paucibacillary: organism numbers are low, the inflammatory response is disproportionate to the bacterial load, and that single fact governs every diagnostic limitation downstream.

Two further mechanistic points separate the consultant from the registrar. First, the disease is overwhelmingly a reactivation phenomenon: a primary infection seeds the tubes during the bacillaemia of childhood or adolescence, then lies dormant within granulomas as latent genital TB until immune control falters, which is why presentation clusters in the reproductive years and why immunosuppression — HIV, but also the iatrogenic immunosuppression of assisted reproduction protocols and the relative tolerance of pregnancy — can tip latent disease into active. Second, the endometrial paradox explains the menstrual picture: because the functional endometrium is shed monthly, granulomas in the superficial layer are often expelled before they caseate, so a single curettage can miss disease that the basal layer harbours — and it is destruction of that basal layer and the underlying vasculature that produces the hypomenorrhoea, amenorrhoea and thin, refractory cavity, not a hormonal failure. The ovary is reached last and least, so ovarian reserve is frequently preserved even when the tubes and cavity are destroyed — a fact that matters enormously when counselling toward IVF, because the eggs are usually retrievable even when the natural pathway is gone.

HIV reshapes gynaecology by a different mechanism — progressive depletion of CD4 T cells removes the cell-mediated immunity that normally contains mucosal and intracellular pathogens, so the changes are patterned and predictable rather than random. The cervix is the organ where this matters most: HIV-positive women acquire HPV more readily, clear it less, and carry multiple high-risk types at once, so CIN incidence runs 2–4× higher and progresses faster to invasive disease, the relationship being bidirectional in that genital inflammation modestly raises HIV shedding. The same loss of containment makes candidiasis recurrent and refractory, herpes simplex lesions severe, atypical and aciclovir-resistant, syphilis more aggressive with a higher neurosyphilis risk, and pelvic inflammatory disease more likely to abscess. The STI–HIV interaction is explicitly bidirectional and is the mechanistic reason syndromic STI control is HIV-prevention work: ulcerative and inflammatory genital infections breach the mucosal barrier, recruit activated CD4 target cells to the genital mucosa and raise both HIV acquisition risk in the negative partner and HIV shedding in the positive one, while HIV in turn worsens the course of every co-infecting organism — a feed-forward loop that an effective treat-and-suppress strategy interrupts from both ends.

Crucially for this objective, immunosuppression is itself a risk factor for TB reactivation, so in the South African setting the two halves of the chapter are not separate diseases but one converging epidemic: a CD4-depleted host both reactivates latent genital TB and tolerates its treatment less well. South Africa carries the highest TB/HIV co-infection burden in the world, with roughly two-thirds of incident TB occurring in HIV-positive people, so the pre-test probability of genital TB in an infertile woman, and the probability that she is also HIV-positive, are both materially higher than the international literature — written largely from the Indian subcontinent — would suggest.

Assessment

Genital TB is found, not suspected, because its symptoms are those of the conditions it mimics. The single most productive context to think of it is the infertility work-up, particularly tubal-factor infertility in a young woman from a high-prevalence community, often with primary infertility and a deceptively normal examination.

• Clinical clues that should raise it: primary or secondary infertility with menstrual disturbance (oligo-/hypomenorrhoea, amenorrhoea, occasionally menorrhagia early on), chronic pelvic pain, a personal or household history of TB, constitutional symptoms, an adnexal mass or ascites, or recurrent implantation failure after otherwise unexplained IVF. A normal pelvic examination does not exclude it.
• Endometrial sampling is the pivotal investigation — premenstrual endometrial aspirate or curettings sent in three parallel streams: histology (caseating granulomas, Langhans giant cells — specific but present in a minority because of the paucibacillary load), mycobacterial culture (the reference standard but slow and insensitive), and GeneXpert MTB/RIF, which simultaneously confirms the organism and reports rifampicin resistance.
• Interpreting GeneXpert is the consultant skill here. On endometrial samples its pooled sensitivity is only about 14% against a specificity near 100% — so a positive result effectively confirms the diagnosis and triggers treatment, while a negative result does not exclude it. The low sensitivity is intrinsic: paucibacillary tissue with few organisms and PCR inhibitors. The practical rule is that you treat on any one positive limb (histology, culture or Xpert) or on a composite clinical-plus-imaging picture, and you do not let a negative Xpert close the question.
• Imaging and endoscopy add the structural diagnosis. Hysterosalpingography shows the beaded or lead-pipe tube, cornual block, calcification and a shrunken or synechiae-deformed cavity; pelvic ultrasound may show a hydrosalpinx, a tubo-ovarian mass or ascites. Laparoscopy with hysteroscopy remains the highest-yield combined assessment — pelvic tubercles, caseous nodules, beaded tubes, peritubal and perihepatic ("violin-string") adhesions and a frozen pelvis are seen, and targeted biopsy and aspiration of peritoneal fluid raise the diagnostic yield — though tubercles regress on therapy while the adhesions and tubal destruction do not, which is the prognostic crux.
• Build a composite, not a single-test, diagnosis. Because no individual test is both sensitive and specific, the working standard is a composite reference: clinical context plus any positive among endometrial histology, culture or GeneXpert, supported by HSG and laparoscopic morphology. A raised CA-125 and ascites in a young infertile woman with a pelvic mass is the classic trap — it mimics ovarian malignancy, and abdominopelvic TB is the benign diagnosis that must be actively excluded before a young woman is committed to cancer surgery; peritoneal-fluid sampling and biopsy at laparoscopy resolve it. The differential also spans chronic PID, endometriosis and Asherman syndrome of instrumental origin, each of which the history and imaging help separate.
• Stage the host and the HIV interface in parallel. Every woman is HIV-tested; in a known-positive woman document the ART regimen and adherence, the current CD4 count and viral load, screen actively for active pulmonary TB (the source focus and a contraindication to immediate elective surgery), and confirm her cervical-screening status — because in South Africa the most likely serious gynaecological lesion in this population is not the TB but a cervical cancer missed through a screening lapse. A chest radiograph and sputum where indicated locate the primary focus and, occasionally, the only culturable source when the genital samples are negative.

Management

Treatment runs on two parallel tracks that share drugs and interactions: curing the mycobacterial infection, and managing the HIV-altered gynaecological host. Organise each as immediate → ongoing → long-term, and hold one honesty throughout — anti-tuberculous therapy reliably sterilises the infection but rarely restores fertility, because it cannot reverse the fibrosis that defines the disease.

Immediate — start standard anti-tuberculous therapy. Genital TB is treated as ordinary drug-susceptible extrapulmonary TB: the WHO and SA NDoH regimen is two months of rifampicin, isoniazid, pyrazinamide and ethambutol (HRZE) followed by four months of rifampicin and isoniazid (HR) — the 2HRZE/4HR six-month course, with pyridoxine cover and rifampicin-resistance excluded by GeneXpert before assuming susceptibility. There is no genital-specific regimen; the diagnosis changes the site you are treating, not the drugs.

Immediate — design around the rifampicin interactions, which are central in this population. Rifampicin is a potent enzyme inducer, and two interactions are non-negotiable in an HIV-co-infected woman of reproductive age:

• ART. Rifampicin lowers dolutegravir exposure, so a woman on first-line TLD (tenofovir–lamivudine–dolutegravir) needs an additional dolutegravir 50 mg twelve hours after her daily dose (effectively dolutegravir twice daily) for the duration of rifampicin and for about two weeks after it stops, because the induction fades slowly. Never let TB therapy quietly undertreat the HIV.
• Hormonal contraception. Rifampicin reduces ethinyl-oestradiol and progestogen levels substantially (oestrogen exposure falls by roughly two-thirds), so combined oral, progestogen-only pill and the subdermal implant become unreliable. The copper or levonorgestrel intrauterine device and depot medroxyprogesterone are the rifampicin-robust choices, and the message must be explicit: this woman can still conceive an ectopic on a damaged tube, so contraception during and shortly after TB treatment is a deliberate decision, not an omission.

Immediate — act on a rifampicin-resistant signal. GeneXpert reports rifampicin resistance with the positive call, and a rifampicin-resistant result converts a routine six-month course into multidrug-resistant TB requiring a specialist all-oral regimen (built around bedaquiline) under the drug-resistant TB programme, with its own longer duration, monitoring and interaction profile. Reading the resistance line, not just the positive/negative, is part of acting on the result.

Ongoing — monitor, support adherence, keep surgery small. Monitor for hepatotoxicity (the HRZ drugs plus a TLD backbone all load the liver) with a low threshold for liver-function testing, optimise nutrition and treat anaemia, and protect adherence through the SA TB programme's directly-observed or self-administered structures, because a six-month course completed is the whole of the cure. In a co-infected woman starting or recently started on ART, anticipate immune reconstitution inflammatory syndrome — paradoxical worsening of TB inflammation as immunity recovers — which can transiently enlarge a pelvic mass or worsen pain and must not be mistaken for treatment failure or malignancy. Surgery has a deliberately limited role — it does not improve fertility and risks injury in a frozen, adherent pelvis — and is reserved for drainage of a persistent tubo-ovarian abscess unresponsive to drugs, a non-resolving mass needing tissue, or symptomatic hydrosalpinges before assisted reproduction. Tubal reconstructive surgery for TB-damaged tubes is futile and abandoned; even the decision to remove a hydrosalpinx is made to improve IVF implantation (a hydrosalpinx leaks embryotoxic fluid into the cavity), not to restore natural patency.

Long-term — counsel fertility honestly and route to IVF, not to the tubes. This is the core of the consultation. After microbiological cure, spontaneous conception is uncommon — the most-cited cohort puts it near 23% conceiving without assistance — and those pregnancies carry a high rate of ectopic and miscarriage because the residual tube is patent but non-functional. The counselling pivots on where the damage sits. Where the tubes are destroyed but the endometrium and ovary are spared, IVF bypasses the problem and offers the best prospect, because retrievable eggs and a receptive cavity are all that is needed. Where the endometrium itself is destroyed by synechiae and a thin, scarred cavity, even IVF struggles: hysteroscopic adhesiolysis is attempted but often disappointing, implantation fails repeatedly, and a markedly damaged cavity is a recognised reason that gestational surrogacy is sometimes the only route to a genetic child. The plan names the cure and the limitation in the same breath: treat the infection, do not promise the fertility, image and assess the cavity before raising IVF expectations, and refer early with realistic counselling — the tube will not be repaired, and the honesty is itself the standard of care.

A short note on menstrual disturbance, which is both a presenting feature and a management problem: hypomenorrhoea or amenorrhoea from endometrial destruction does not respond to hormones (the target tissue is scarred, not under-stimulated), so the response is to characterise the cavity hysteroscopically and counsel rather than to escalate oestrogen; the occasional early heavy bleeding settles as inflammation resolves on therapy.

The HIV-gynaecology half — manage the host the TB sits in. The general principles of HIV-altered gynaecology are covered at registrar level; the consultant additions are about intensity and integration.

• ART optimisation is gynaecological treatment. A suppressed viral load and a recovering CD4 count improve every co-morbid problem — they reduce the frequency of candida, BV and HSV recurrences, slow CIN progression, and improve TB and surgical tolerance — so confirming suppression and correcting non-adherence is the first management step, not a referral afterthought. The framing throughout is PVT-era HIV care (prevention of vertical transmission), not the older PMTCT language.
• Intensified cervical screening is the highest-value single act. Per SA policy, women living with HIV are screened at HIV diagnosis and then every three years (more often if abnormal), independent of age, with a lower threshold for colposcopy and excisional rather than ablative treatment of CIN2/3 given the higher recurrence with low CD4 and unsuppressed virus. A screening lapse here is the commonest preventable cervical-cancer death in South African gynaecology, and the move toward primary HPV testing — more sensitive, with self-sampling that can lift coverage in exactly this hard-to-reach population — is the system-level intervention with the most to offer it.
• Use the encounter as an integration point. The practical SA lesson is that the woman in front of you may attend several vertical programmes — HIV, TB, antenatal, family planning, cervical screening — and the gynaecological consultation is the moment to close the loops the system leaves open: confirm she is on ART and suppressed, that any TB is treated, that she is screened and contracepting safely on a rifampicin-robust method, and that her partner and children are linked to care. Treating the visit as a single integrated encounter rather than a referral relay is what converts good guidelines into delivered outcomes.
• Treat the amplified common infections as chronic problems. Recurrent vulvovaginal candidiasis and frequent herpes recurrences often need suppressive rather than episodic therapy — long-course or maintenance azole for refractory candidiasis, daily suppressive aciclovir/valaciclovir where recurrences are frequent — and severe or aciclovir-refractory genital herpes needs dose escalation and, if it persists, a search for resistance and a switch to foscarnet under specialist care. Bacterial vaginosis is more common and contributes to PID and upper-tract risk, so it is treated rather than tolerated. PID in an HIV-positive woman warrants a lower threshold for admission and intravenous therapy, a higher index of suspicion for tubo-ovarian abscess, and active consideration of TB-PID when the picture is chronic, indolent or atypical rather than the acute polymicrobial presentation — genital tuberculosis arriving in the clinic dressed as something else.
• Address menstrual and fertility consequences of HIV in their own right. Beyond TB, HIV is associated with a higher rate of premature ovarian insufficiency and hypothalamic menstrual disturbance, much of which improves with immune reconstitution on ART; in serodiscordant couples wanting to conceive, U=U applies — an undetectable, suppressed partner does not transmit, so safe natural conception is possible without sperm-washing, and the consultation is about timing intercourse to suppression rather than about technology.
• Plan elective surgery around immune reserve. HIV status alone does not contraindicate gynaecological surgery; a CD4 above 200 with a suppressed viral load gives equivalent outcomes, while a low CD4 or detectable virus raises infective and wound complications and argues for optimisation first. Screen for and treat active TB before any elective operation, give standard surgical antibiotic and VTE prophylaxis (HIV modestly raises thrombotic risk), and run a deliberate drug-interaction check — rifampicin, ritonavir-boosted regimens, azoles, ergometrine and statins all interact.
• Do not strip out the psychosocial dimension. Depression affects a large minority of women living with HIV, the infection is frequently acquired through sexual violence, and disclosure and stigma shape adherence; routine mental-health screening and referral are part of competent management, not an optional extra, and in adolescents — including a growing cohort of vertical-transmission survivors — confidentiality and disclosure history change the whole consultation.

Guidelines compared

The instructive feature is not disagreement but the division of labour and a terminology gap: the TB and HIV programmes are governed by separate documents that the gynaecologist must integrate, and the SA framing has moved ahead of older international language.

The genuine practice tension is not between guidelines but between guideline and service reality — GeneXpert is available but insensitive, ART and TB drugs are free but adherence over six months is the limiting step, and the screening interval is policy but coverage is the failure point.

The evidence & the controversy

The evidence base for genital TB is cohort-led, not trial-led, and a candidate must say so rather than overstate it: there are no randomised trials defining its therapy, because the condition is too uncommon in high-resource settings to power one and too clearly TB to ethically withhold treatment in endemic ones. What exists is observational. The most-cited fertility figure — that roughly one in five treated women conceives without assisted reproduction, with a disproportionate share of those pregnancies ending in ectopic or miscarriage — comes from infertility-clinic cohorts, not population data, and almost certainly overstates the real-world yield because it is drawn from women fit enough to reach a tertiary fertility service. The honest consultant position is that anti-tuberculous therapy is a microbiological success and a fertility disappointment, and that the central decision is to set that expectation early and route to IVF rather than to repeated surgery on dead tubes.

The diagnostic controversy is the GeneXpert paradox. A test with near-perfect specificity but ~14% sensitivity on endometrial tissue is genuinely useful for ruling in and genuinely dangerous if a negative is read as a rule-out. This drives a quiet but important real-world practice — empirical anti-tuberculous treatment on a strong composite picture (clinical, HSG and laparoscopic) despite negative molecular confirmation — which sits uncomfortably between under-diagnosing a treatable cause of infertility and over-treating women with six months of hepatotoxic drugs they may not need. There is no clean resolution; the defensible stance is to demand at least a coherent composite (imaging plus clinical context, ideally with histology or laparoscopy) before committing, and to be candid that the diagnosis is sometimes probabilistic.

A current thread worth holding is the rising recognition of latent and reactivation genital TB in assisted-reproduction populations. As IVF access widens, recurrent unexplained implantation failure with a thin or scarred endometrium is increasingly being investigated for occult genital TB, and there is a real and unresolved debate about whether and how to screen for latent genital TB before embryo transfer in endemic settings — the risk being both a missed treatable cause and the over-treatment of women with six months of toxic drugs on weak evidence. The broader South African anxiety is drug-resistant TB: a rifampicin-resistance signal on GeneXpert turns a routine six-month problem into a multidrug-resistant one requiring specialist regimens, which is precisely why the resistance read-out, not just the positive call, matters at the point of diagnosis — and it is one reason the recent national progress in cutting TB incidence and the move toward shorter, all-oral and BPaL-class regimens are watched closely even from the gynaecology clinic.

The HIV half carries its own live controversies. The long-running question of whether injectable progestogen contraception affects HIV acquisition has now largely settled toward "no major increase," which keeps DMPA — a rifampicin-robust, widely used SA method — firmly on the menu rather than restricted. More pointedly, the cervical-cancer burden in women living with HIV is a screening-coverage failure as much as a biological one: the policy of screening at diagnosis then three-yearly is sound, but the gap between policy and delivered coverage is where preventable cancers arise, which keeps prevention and system-level follow-up, not just individual treatment, squarely in the consultant's remit. The defensible viva position across both halves is the same — name the SA-specific protocol precisely, be candid that the evidence underpinning genital-TB management is observational rather than randomised, and let the resource and coverage realities, not just the biology, shape the plan.

Landmark trials & key evidence

Genital TB has no defining randomised trials; the evidence that grounds its management is the canonical narrative review, the diagnostic meta-analysis, and the fertility-outcome cohorts, alongside the WHO regimen and the SA epidemiological frame. The absence of RCT-level evidence is a feature of the field to acknowledge openly rather than a gap to paper over.

Exam traps & red flags

• Reading a negative GeneXpert as a rule-out. With ~14% sensitivity on endometrium, a negative Xpert excludes nothing; treat on any positive limb or a coherent composite, and never tell an infertile woman she "does not have TB" on a single negative molecular test.
• Promising fertility from anti-TB therapy. The drugs cure the infection; they do not rebuild fibrosed tubes or a synechiated cavity. Counselling that implies natural conception will follow is the central error — the honest message is microbiological cure with poor spontaneous fertility and early referral for IVF.
• Operating to restore fertility. Tubal reconstruction in TB is futile and the pelvis is dangerous to enter; surgery is for abscess drainage or tissue, not for tubal repair.
• Forgetting the rifampicin–dolutegravir interaction. Leaving a co-infected woman on once-daily dolutegravir while she takes rifampicin undertreats her HIV; the additional 50 mg dose (effectively twice daily, continued ~2 weeks after rifampicin) is mandatory.
• Relying on the pill or implant during rifampicin. Enzyme induction makes oestrogen- and progestogen-based oral methods and the implant unreliable; default to the IUD or DMPA, and remember a damaged tube can still carry an ectopic.
• Missing rifampicin resistance on the GeneXpert read-out. A rifampicin-resistant signal converts a routine six-month course into an MDR-TB problem needing a specialist regimen — read the resistance line, not just the positive call.
• A cervical-screening lapse in a woman living with HIV. The most likely lethal gynaecological lesion in this population is cervical cancer missed through a skipped screen; WLHIV are screened at diagnosis then 3-yearly, not on the age-30 general schedule.
• Anchoring on recurrent candida without testing for HIV. Refractory or recurrent thrush, severe atypical herpes, or unusually severe PID should prompt an HIV test — the gynaecological pattern is often the presenting clue.
• Operating electively without staging the host. A low CD4, an unsuppressed viral load or untreated active TB each raise complication risk; optimise ART, exclude active TB, and run the drug-interaction check before scheduling.
• Using PMTCT language and old guideline framing. SA practice is PVT-era; the terminology has changed and the dolutegravir-doubling rule on rifampicin is the practical consequence.

Evidence anchors

• Sharma JB et al. — Female genital tuberculosis: Revisited, Indian J Med Res 2018;148(Suppl 1):S71–S83
• Palo S et al. — Performance of pelvic-derived samples for diagnosis of female genital TB: systematic review & meta-analysis, Cureus 2025
• Kulshrestha V et al. — Genital TB among infertile women and fertility outcome after anti-TB therapy, Int J Gynaecol Obstet 2011;113:229–34
• WHO consolidated guidelines on tuberculosis, Module 4: drug-susceptible TB treatment (2022)
• WHO Global Tuberculosis Report 2024
• South Africa NDoH National Consolidated Guidelines for the Management of HIV — TLD first-line; rifampicin co-treatment requires dolutegravir 50 mg twice daily; PVT-era terminology.
• South Africa NDoH Cervical Cancer Prevention and Control Policy — women living with HIV screened at HIV diagnosis then every 3 years; HIV-negative women from age 30; transitioning toward primary HPV testing.
• Intermediate groundwork: HIV and gynaecological health — the patterned effects of HIV on cervical, vulvovaginal, pelvic and menstrual disease assumed here.