Trachoma

Trachoma is a neglected tropical disease caused by infection with conjunctival strains of Chlamydia trachomatis. It can result in blindness. Pathophysiologically, trachoma is a disease complex composed of two linked chronic processes: a recurrent, generally subclinical infectious–inflammatory disease that mostly affects children, and a non-communicable, cicatricial and, owing to trichiasis, eventually blinding disease that supervenes in some individuals later in life. At least 150 infection episodes over an individual’s lifetime are needed to precipitate trichiasis; thus, opportunity exists for a just global health system to intervene to prevent trachomatous blindness. Trachoma is found at highest prevalence in the poorest communities of low-income countries, particularly in sub-Saharan Africa; in June 2021, 1.8 million people worldwide were going blind from the disease. Blindness attributable to trachoma can appear in communities many years after conjunctival C. trachomatis transmission has waned or ceased; therefore, the two linked disease processes require distinct clinical and public health responses. Surgery is offered to individuals with trichiasis and antibiotic mass drug administration and interventions to stimulate facial cleanliness and environmental improvement are designed to reduce infection prevalence and transmission. Together, these interventions comprise the SAFE strategy, which is achieving considerable success. Although much work remains, a continuing public health problem from trachoma in the year 2030 will be difficult for the world to excuse. Trachoma, caused by infection with conjunctival strains of Chlamydia trachomatis, is the most common infectious cause of blindness. This Primer summarizes the epidemiology, pathophysiology and diagnosis of trachoma as well as its management, disease control and elimination, and key areas for future research.

Trachoma is one of 20 diseases and disease groups desig nated as neglected tropical diseases by the World Health Organization (WHO) 1 . It is the most common infectious cause of blindness 2 . The disease can be conceptualized as progressing in two phases. In the first phase, repeated infection 3,4 with conjunctival strains 5 of the bacterium Chlamydia trachomatis (Ct) results in a chronic kerato conjunctivitis (inflammation of the cornea and con junctiva), including an inflammation of the conjunctiva known as active trachoma. Rounds of severe conjunctival inflammation lead to scarring of the eyelid 6,7 , which is the start of the second disease phase. This scarring can cause inward rotation of the eyelashes so that they come into contact with the eyeball, a condition known as tri chiasis. Trichiasis may be accompanied by a distortion of the eyelid known as entropion 8 , in which part or all of the eyelid margin is rolled inward. Scratching of the cornea by inturned eyelashes predisposes to corneal opacity, vision impairment and blindness. The presence of any of these Ctinduced pathological processes or clinical signs is trachoma. However, to fully grasp the nature of this disease, it is imperative to also consider the devastating effects of trichiasis and vision impairment on affected individuals, their families and their communities 9 . In economic terms alone, these effects combine to reduce global productivity by billions of dollars each year 10 .
Trachoma is a disease of poverty -it affects the poorest of the poor 11 . In Europe and most of North America, trachoma disappeared decades ago as living standards improved and without the implementation of specific interventions 12 . Similar trends have been seen in some lowincome and middleincome countries with previously hyperendemic disease [13][14][15] . In addition, during the past few decades, deliberate interventions to control trachoma have been associated with dramatic declines in its global burden 16,17 . However, prevalence has remained high in some populations despite prolonged intensive intervention with a comprehensive, fourcomponent strategy recommended by WHO 18 .
In this Primer, we provide an overview of the epide miology, pathophysiology, clinical features and diagno sis of trachoma as well as of the management, control and elimination of disease, at both the individual and population level. We also propose key areas for future research.

Epidemiology
Global trachoma prevalence data are published annu ally by WHO 17 and summarized on the WHO Global Health Observatory. Data at a finer scale are available on the Trachoma Atlas. As of March 2022, 44 countries were known to require interventions against trachoma, of which 26 were in the WHO African Region.

Associations
Poverty. At any spatial scale at which observations are made, trachoma is found in the poorest people. Risk of active trachoma is higher in households with crowded sleeping arrangements 11,[19][20][21] ; sharing a bedroom with someone else who has active trachoma doubles an indi vidual's risk 22 . In trachomaendemic communities, individuals with trachomatous trichiasis (TT) are likely to be poorer than agematched and gendermatched peers without TT and are less likely than those peers to participate in economically productive activities even after controlling for visual impairment 11 . In trachoma endemic countries, the disease affects the population groups that are poorest, most marginalized and most remote from services (particularly water and sanita tion) 23 . At the regional and global level, concentration of trachoma in the poorest countries is evident 17,24 .
Poverty is both a cause and consequence of blindness 25 . Trachoma can trap affected individuals, families and communities in successive generations of despair.
Age. In trachomaendemic populations, individuals are often first exposed to conjunctival Ct during the first months of life 26 . Ct infection and active trachoma occur most commonly during childhood; in infected individuals, Ct loads are higher at younger ages [26][27][28][29][30][31] . In hyperendemic areas, the frequencies with which infection and active trachoma are observed decrease with age after a peak in those aged 2-5 years. TT, corneal opacity and visual impairment are unusual before adulthood and observed frequencies increase with age 27,32 .
Gender. In childhood, boys and girls tend to be equally affected by active trachoma 33,34 . For TT, in pooled data, women have 1.8 times higher odds of being affected than men 35 , although this odds ratio can approach 4 in some settings 33,36 . The excess risk is generally attributed to women being exposed to conjunctival Ct more fre quently than men as they get older because of women's disproportionate contribution to childcare duties in most societies 37 . More direct biological effects of oes trogen, oestrogen receptor abundance or sexlinked dif ferences in the immune response to Ct 38 have not been ruled out.

Mechanisms/pathophysiology Chlamydia trachomatis
Ct is a gramnegative bacterium that infects humans at the epithelial layer of mucosal surfaces. Different Ct serovars, originally distinguished using micro immunofluorescence 39 , have a characteristic tissue tropism and are associated with different disease complexes: A, B, Ba and C with trachoma, D through K with urogenital chlamydia, and L1, L2 and L3 with lymphogranuloma venereum (Box 1).
Ct has a unique biphasic developmental cycle 40 . It moves between host cells as the infectious form, called the elementary body (diameter ~0.30 μm). Using mostly receptormediated endocytosis, the elementary body gains entry to a human epithelial cell within a vacuole formed of host cell membrane and chlamydial proteins, which then develops into a perinuclear inclusion 41 . Inside the inclusion, it reorganizes into its more metab olically active, replicating, noninfectious form, the reticulate body (diameter ~1.2 μm). After ~72 h, the retic ulate body transforms into 100-1,000 elementary bodies that are released via host cell lysis. In the intracellular environment, Ct evades detection and dampens the host immune response via several mechanisms 40 .
The elementary body has a rigid cell wall, which facil itates survival under unfavourable conditions. The major outer membrane protein is the most abundant cell wall surface protein, comprising ~60% of cell wall mass. It is involved in host cell surface adhesion and is encoded by a singlecopy chromosomal gene, OmpA 42 . Variation in this protein characterizes the 19 different Ct serovars. Wholegenome sequencing revealed that the preference of different serovars for particular anatomical niches is not absolute. Extensive recombination in the OmpA region has resulted in exchange of genetic material both within and between conjunctival, urogenital and lym phogranuloma venereum strains 43 . Urogenital strains are generally able to synthesize their own tryptophan, whereas conjunctival strains typically have inactivating mutations in the gene encoding tryptophan synthase 44 . One possible explanation is that ocular serovars are able to access a source of tryptophan (or one of its biosyn thetic precursors) within the conjunctival sac or infected cells, and that that source is unavailable to pathogens in the urogenital tract 44 . Of note, tryptophan itself is not found in human tears 45 . The mechanism by which conjunctival Ct survives in this milieu and whether its peculiar biochemistry could be exploited against human disease is unknown.
Other Ct proteins in addition to the major outer membrane protein are produced in a complex sequence at different stages in the developmental cycle 46 ; these include heat shock proteins that protect Ct in stress con ditions. These proteins, such as cHsp60, are implicated in disease pathogenesis and are recognized by patho gen recognition receptors of the innate immune system, including Tolllike receptors 47 .

Transmission of Ct
Transmission of conjunctival Ct infection has been the subject of considerable scrutiny, yet methods to directly investigate transmission routes in trachomaendemic communities were not applied until 2018. Important epidemiological evidence used to generate biologically plausible hypotheses on transmission includes the focal nature of infection and disease, with spatial clustering of cases identified at bedroom, household, compound, neighbourhood and community levels 31,[48][49][50][51][52][53][54][55][56] ; the asso ciation of active trachoma with visible eye and nose discharge on children's faces [57][58][59] ; and the association of active trachoma with observed flyeye contact 53,60-62 . All of the above tend to be observed in populations that are relatively overcrowded 57,62,63 and have poor access to water and sanitation 59,63-66 . Infected eye discharge or cul tured Ct can induce active trachoma when directly inoc ulated into an eye 19,67 , suggesting that simple mechanical transfer of infectious material from an infected to an uninfected eye may be all that is necessary to create a transmission event. Because tears drain into the nose and Ct can be identified in the noses of children with active trachoma 68,69 , nasal discharge may also be impor tant in transmission of infection. Together, these obser vations implicate three principal routes of transmission: direct persontoperson transfer of infected secretions on human fingers; spread via fomites such as shared bed ding or towels; and carriage on eyeseeking flies, particu larly female Musca sorbens, which preferentially obtain protein from human exudates for egg production 70 . It is not yet known with certainty whether Ct infects and replicates within vector fly species but laboratorybased work on flies from the same genus 71 shows that viable Ct can be retrieved from flies for up to 48 h after feeding. Fingers, fomites and flies are sometimes referred to as the 'three Fs' of trachoma transmission.
Evidence now substantiates these proposed routes. Ct DNA can be detected on flies caught leaving the faces of children in trachomaendemic communities 55,70,72,73 . In Ethiopian households in which no residents had current conjunctival Ct infection and which were located close to other households in which one or more residents did have conjunctival Ct infection, some flies caught leaving the faces of children were positive for Ct DNA. This impli cates flies as potential carriers of conjunctival Ct between households as well as within them 55 . In a prospective trial, fly control with insecticide space spraying, which involves the creation of a liquid fog throughout commu nity outdoor spaces, although not seen to be a sustaina ble intervention at scale, reduced the prevalence of active trachoma 74 , confirming the importance of flies as vectors.
A study published in 2020 demonstrated the pres ence of Ct DNA at other extraocular sites: on the skin of human faces and hands, clothing, a sleeping surface, and a washing jug in households in which one or more res idents had conjunctival Ct infection 55 . In households in which all residents had Ctnegative conjunctival swabs, none of these extraocular sites had detectable Ct DNA 55 . Experiments employing viability PCR, which uses propidium monoazide pretreatment to prevent ampli fication of DNA from nonviable bacteria, suggest that at least some Ct remains viable on plastic, cotton cloth and skin for >24 h (ref. 75 ), providing additional support for the roles of fingers and fomites in transmission.

Conjunctival Ct infection and active trachoma
Epidemiological observations have informed mathe matical models of trachoma, which suggest that >100 conjunctival Ct infections in an individual's lifetime are required to generate clinically significant conjunctival scarring and that 150 infections are required to precipi tate TT 76 . As most episodes of infection occur in child hood, children who go on to develop TT later in life are likely to be reinfected several times each month, with transfer of Ct back and forth from one child to another.
The limited longitudinal data available indicate that a primary infection episode in a child is character ized by a short 'preclinical' phase of a few days before the development of signs of inflammation 77 . In adult human volunteer experiments, clinical inflammation developed ~10 days after conjunctival Ct inoculation 78 . Established infections are characterized, particularly in younger preschool age children, by a generalized folli cular and papillary conjunctivitis, particularly evident in the conjunctiva of the upper eyelid. At the tissue level, the generalized papillary inflammation involves a mixed inflammatory cell infiltrate, including lymphocytes, macrophages and neutrophils. Lymphoid follicles are formed mostly of B cell aggregations in the conjunctival stroma. After the development of clinical inflammation, infection remains detectable for a few days to many weeks 77 . The human immune response then controls the infection, clearing it or reducing it to undetectable levels. However, conjunctival inflammation persists, and may last for many weeks after infection becomes unde tectable; in untreated children aged 4-15 years, infection may be cleared after 3-8 weeks, whereas clinical signs of

Box 1 | Disease associations of non-ocular Ct serovars
Urogenital Chlamydia trachomatis (Ct) infection is sexually transmitted, causing urethritis, cervicitis, proctitis, epididymitis and pelvic inflammatory disease. In some women, scarring of the fallopian tubes occurs, presumably from repeated infections, leading to infertility or ectopic pregnancy. Lymphogranuloma venereum is also sexually transmitted, with Ct entering via the mucosae, affecting local lymphatic vessels and lymph nodes, and occasionally becoming disseminated 292 . An increasing body of both experimental and observational work challenges the notion that only lymphogranuloma venereum Ct strains invade beyond the epithelium of the tissue of entry, indicating that serovars A-K, which cause trachoma and urogenital infections, probably also have this ability 293,294 . inflammation continue for 6-18 weeks 79 . Resolution of infection is thought to depend on a cellmediated immune response, effected through IFNγ 80 . Adults tend to have shorter episodes of infection and disease, although the relative frequency of exposure to reinfec tion in children and adults cannot be quantified 77 . This protection in older individuals is thought to be medi ated through acquired T celldependent cellmediated responses that either prevent or more rapidly resolve infection 80 .
The different time courses of conjunctival infec tion and active trachoma explain the partial mismatch observed between infection and disease at the individ ual level 54,[81][82][83] . This disparity has important implications for trachoma programmes, as individually targeted antibiotic treatment decisions based on observable conjunctival inflammation of at least moderate inten sity would miss many infected individuals with a mild clinical response. Antibiotic mass drug administration (MDA) is the most practical way of trying to treat all individuals infected with conjunctival Ct within large trachomaendemic populations 84 .

Development of scarring sequelae
The association between increasing age and increasing prevalence of cicatricial sequelae of trachoma (conjunc tival scarring, TT and corneal opacity) indicate that these signs are cumulative. Several longitudinal studies have investigated their natural history and pathophysiology 85 . To date, five studies have examined the rates and risk factors for progression from conjunctivae without scar ring through active trachoma to the development of con junctival scarring 6,7,[86][87][88] . Their findings show a consistent and substantially increased risk of subsequent incident scarring associated with intense conjunctival papillary inflammation, particularly when observed at multiple time points 6,7,86,88 . Evidence linking subsequent scar ring to the presence of a follicular conjunctivitis (with out concomitant intense papillary inflammation) or to repeated or constant Ct infection is much less convinc ing. Only two published cohort studies of incident scar ring have prospectively tested for Ct infection. Neither of these has identified a relationship between ongoing Ct infection and incident or progressive scarring after adjusting for clinical inflammation 7,88 .
Two large prospective studies conducted in Ethiopia and the United Republic of Tanzania have followed the progression of preexisting conjunctival scarring in adults, including tests for the presence of Ct 89 . Both found strong evidence of a relationship between progres sion of scarring and the repeated observation of papillary inflammation of the conjunctiva. Conjunctival Ct infec tion was very rare in these cohorts, and the infections that were detected were not associated with progres sion. However, it is possible that study participants were infected occasionally for very short periods 77 , which multiple crosssectional samples could easily miss.
Taken together, these studies point to a central role for chronic severe inflammation in scarring develop ment ( fig. 1). Such severe responses are not found in all people exposed to conjunctival Ct, indicating that varia tions in the human immune response may be important codeterminants of scarring risk although it is not possible to quantify individual exposure to repeated reinfection. At the population level, a higher prevalence of inflammatory trachoma in children correlates with a higher prevalence of trichiasis in adults 90 . Available data also suggest that additional proinflammatory stimuli, such as other bacteria and ocular surface dryness 81 , contribute to the progression of scarring initiated by Ct-related inflammation 91 .

Immunopathological basis of scarring trachoma
Trachoma offers an unusual opportunity to directly observe and investigate the pathophysiology of an inflammatory scarring disease process, with poten tial relevance to disease processes elsewhere. Multiple longterm natural history studies have been con ducted, applying a range of techniques (histopathology, immunohistochemistry, in vivo confocal microscopy, human gene expression profiling or human genetics) and relating these to the clinical course 80 . In addition, nonhuman primate models have been developed as part of Ct vaccine development work, which have con tributed substantially to our understanding of trachoma pathophysiology 3,92-94 .
Biopsy and in vivo confocal microscopy studies of children with active trachoma demonstrate that it is characterized by a mixed conjunctival inflammatory cell infiltrate, including macrophages, T cells, neutrophils and dendritic cells 95,96 . Organized lymphoid follicles are scattered throughout the conjunctival stroma, composed largely of B cells surrounded by a mantle of proliferating lymphocytes ( fig. 2).
In biopsy specimens from adults with conjuncti val scarring, extensive connective tissue disruption can occur with loss of the regular stromal architec ture and replacement with disorganized collagen and other elements 97 . Compared with controls, scarred tis sue also has marked increases in CD45 + inflammatory cells, whose abundance fluctuates with the intensity of clinical inflammation 98 . Specific staining indicates that natural killer cells are a prominent component of the cellular infiltrate in scarred conjunctivae, suggesting that innate responses may be relevant in this damaged tissue 98 . Staining of scarred conjunctivae against a panel of cytokines and scarring disease markers reveals prom inent increases in epithelial expression of connective tissue growth factor (CTGF) and the antimicrobial pep tide S100A7; in addition, expression levels of IL1β are increased in the substantia propria 99 .
In vitro infection studies of epithelial cells have found a marked innate proinflammatory response, with the production of several cytokines: IL6, IL8, growthregulated oncogeneα (GROα) and granulocytemacrophage colonystimulating factor (GMCSF) 100,101 ( fig. 3). This finding is consistent with data from in vivo investigations into conjunctival gene expression profiles using swabs taken from the conjunctival surface [102][103][104][105] . The initial innate response to Ct is likely to be driven directly by infected epithelial cells, recognizing the pres ence of the organism through their patternrecognition receptors, leading to a chronic inflammatory response. This suggestion is sometimes referred to as the cellular paradigm for the pathogenesis of Ctdriven scarring disease 106 . The alternative hypothesis, that is, the immuno logical paradigm, suggests that inflammation and tissue damage are driven through a cellmediated immune or delayedtype hypersensitive response to chlamydial antigens such as cHsp60. Although well established in animal models, direct evidence from human studies corroborating this is limited 80,94,106 .
Resolution of Ct infection is thought to be depend ent on IFNγ from a CD4 + T helper 1 (T H 1) cellmediated immune response. Gene expression studies in children in trachomaendemic populations show a marked increase in conjunctival expression of IFNγ and T H 1 cellrelated factors (IL12B and indoleamine2,3dioxygenase (IDO1)) and in natural killer cell pathways (characterized by NCR1, CD56 and CD247) [102][103][104] . In children with active trachoma but no detectable Ct, levels of IFNγ are not par ticularly elevated, suggesting that this response is rapidly regulated following the resolution of infection. Infection and active trachoma are also both associated with profiles that are consistent with a prominent T H 17 response, with increased expression of IL17, IL21 and IL22 (ref. 104 ). Some data indicate that T H 17-IL17 responses might contribute to a worsening of the inflammatory/scarring response in animal models 107 .
Intense conjunctival inflammation is the key clin ical manifestation linked with the development of scarring. In crosssectional studies, intense inflamma tion is associated with increased expression of a range of proinflammatory factors such as S100A7, defen sin (DEFB4A), IL1β, IL17A, CCL18 and neutrophil chemotactic factor CXCL5 (ref. 104 ). These often seem to persist after Ct infection has resolved. Interestingly, several markers (S100A7, IL1β, IL17A and CXCL5) are also consistently elevated in adults with established conjunctival scarring and visible inflammation 89 .
In the context of this chronic inflammatory milieu, with the recruitment and activation of a diverse pop ulation of leukocytes, the conjunctival tissue is repeat edly damaged and scar tissue is formed during healing. Matrix metalloproteinases (MMPs) are a large and diverse family of proteases that are central to the regu lation of connective tissue in health and disease. Studies in trachomaendemic populations have found increased expression of MMP7, MMP9 and MMP12 in children with intense conjunctival inflammation and adults who have established scarring with inflammation 89,102,104,108 . In a cohort study of Tanzanian children assessed every 3 months for 4 years, progressive scarring was strongly associated with increased proportions of followup points at which clinical inflammation was seen 88 as well as with increased expression of proinflammatory chemokines (CXCL5, CCL20, CXCL13 and CCL18), cytokines (IL23A, IL19 and IL1B), MMP12, and S100A7 and reduced expression of SPARCL1 (ref. 105 ).
Increased expression of several fibrogenic growth factors, including CTGF, FGF, TGFβ1 and PDGF, has also been associated with clinical inflammation in active trachoma and scarring 89,104,108 . These factors have a profibrotic effect, probably largely mediated through increased deposition of connective tissue elements by stromal fibroblasts. Interestingly, cultured conjunctival fibroblasts from individuals with trachomatous scar ring have a markedly contractile profibrotic phenotype compared with control fibroblasts, suggesting some (asyet uncharacterized) permanent change in their behaviour 109 .

Diagnosis, screening and prevention Diagnosis
Trachoma is diagnosed through clinical examination. This requires a trained examiner, a calm examinee, mag nification (generally ×2.5 binocular magnifying loupes), and a means for the examiner to prevent iatrogenic transmission of infection from one examinee to the next  7) Surgery reduces progressive corneal opacification and ocular surface irritation, which is crucial for symptom relief. Surgery has no effect to correct existing ocular surface dryness. (8) In many conditions that blind through corneal opacification, keratoplasty (corneal transplantation) is used to reduce or reverse visual impairment. In an individual with corneal opacity due to trachoma, changes in the tear film and corneal vascularization are predicted to reduce the likelihood of long-term graft survival. In addition, populations affected by trachoma tend to have very poor access to keratoplasty services.
(alcoholbased hand gel, disposable examination gloves, or soap and water). Diagnosis is often (but not always) done in the field. Illumination is also important; a torch is optimal when looking for TT, whereas bright sunlight is generally adequate for examining the conjunctiva. Most individuals with active trachoma have min imal or no symptoms. Some report eye irritation or a small amount of discharge. Individuals with TT often report a history of eye pain, blepharospasm (involun tary closure of both eyelids), light intolerance or reduced vision 110 . Many epilate their inturned eyelashes using homemade forceps, clam shells or tree resin 111 . Poor vision may be reported even in the absence of corneal opacity. Quality of life may be severely impaired 9,112,113 .
A comprehensive trachoma examination routine is ideal but, in some contexts (in particular, routine sur veys conducted by public health programmes to esti mate prevalence), not all of its elements are required. The examination should be guided by its purpose and context.
The comprehensive examination comprises several steps. Visual acuity is measured (for each eye separately) where possible. To reduce the risk of transposing find ings in the record, routinely examining the right eye, then the left eye, is recommended. Upper and lower eyelids are inspected for entropion, trichiasis and evidence of recent epilation; the latter includes broken or regrowing eyelashes or empty eyelash follicles. The cicatricial com plications of trachoma characteristically affect the upper eyelid; in ~10% of patients, they also affect the lower eye lid but rarely in the absence of upper eyelid disease 114 . The cornea and limbus (border between cornea and sclera) are inspected for opacities, upper pole pannus (ingrowth of fibrovascular tissue from the limbus into the cornea) and Herbert pits (rounded depressions at the limbus resulting from regression of lymphoid follicles). If fluorescein and a slit lamp are available, superficial punctate keratitis might be detectable in some individuals with active trachoma. The upper eyelid is everted and the tarsal conjunctiva examined for central follicles, inflammatory thickening due to a papillary response, and scarring 110,115 . If follicles are present in the central part of the upper tarsal con junctiva, only those with a diameter ≥0.5 mm should be regarded as being of pathological significance; follicle size guides 116 fixed to the examiner's thumbnails can assist in accurately assessing follicle diameter. Particularly when examining children, having an assistant to reassure and support the examinee can be invaluable 117 .  Other causes of conjunctival inflammation, conjunc tival scarring, trichiasis, entropion and corneal opacity can have clinical signs that resemble those produced by trachoma. For example, inclusion conjunctivitis looks like active trachoma but is a selflimited inflammatory condition caused by infection of the conjunctiva with urogenital Ct 118 . Its prevalence has not been determined for any trachomaendemic population but could be high in places where urogenital Ct is very common and access to water and sanitation is very poor. Nonchlamydial bacterial or viral infection may also drive follicular inflammation of the conjuctiva 81,119 .
The interpretation of clinical signs and their clas sification as being attributable (or not attributable) to trachoma can be influenced by a range of contextual impressions. For trachoma, these include the rela tive poverty of the patient and family, the examiner's impression of the patient's personal and family hygiene, and the trachoma endemicity of the patient's commu nity. However, communitylevel endemicity status is often unknown or not known with certainty. The examiner's view on whether trachoma is endemic in a community influences the clinical interpretation of the patient, which can then (potentially inappropriately) confirm, in a circular fashion, the examiner's classifi cation of the community 120 . This possible confirmation bias is important in both the diagnosis of active tra choma and in differentiating upper eyelid trichiasis that is trachomatous from that caused by other pathological processes.
To help maximize the objectivity of assessment, par ticularly for estimates of trachoma prevalence at the population level, WHO developed trachoma grading systems that have subsequently evolved over decades 110 . Two systems are in current use. The modified WHO grading system (or FPC system, for follicles (F), pap illary hypertrophy (P), cicatriciae (C)) is a specialist scale that was last updated in 1981 (ref. 121 ) and is used principally in research studies. The simplified WHO grading system, intended for nonspecialist personnel, was first published in 1987 (ref. 122 ) and updated in 2020 (ref. 115 ) ( TaBle 1). It is in widespread use for screening and populationlevel assessment. The five signs are signs, not stages, and more than one can be simultaneously present in an eye. In the simplified grading system, the presence of trachomatous inflammation -follicular (TF) and/or trachomatous inflammation -intense (TI) in either eye constitutes active trachoma ( fig. 4).
The simplified system includes acknowledged imperfections 115 . For example, not all epilation is under taken for trichiasis 123 ; when epilated eyelashes have not fully regrown, it is hard to determine whether they were touching the eyeball before removal; and not all upper eyelid trichiasis is due to trachoma 124 .
Since 2012, standardization of grader training, grader certification and fieldwork procedures 116,[125][126][127][128] have further contributed to comparability and confidence in trachoma prevalence data generated by national programmes. However, when assessing an individual patient, rigid adherence to grading systems and standard operating procedures may not be appropriate. This rep resents a fundamental difference between the diagnostic approaches needed for clinical (Box 2) and epidemiological (Box 3) purposes.

Inflammatory cell infiltration
Dead cell

Pro-inflammatory epithelium Resolution of infection C. trachomatis infection
Lymphoid layer

Screening
In June 2021, 136 million people worldwide lived in areas in which active trachoma was thought to represent a public health problem 17 . Individually screening large populations for active trachoma is problematic because examining millions of people requires many examin ers, and examiners without adequate training provide assessments with predictive values that are too low to be epidemiologically or programmatically useful 129 . Even physicians working in endemic areas may produce inaccurate screening assessments if they are not well trained 130,131 . Of note, the frequency of asymptomatic Ct infection and ease of reinfection after treatment mean that characterization and management of the population as a whole are crucial for controlling active trachoma; individual screening for the purposes of individuallevel management would be unhelpful for achieving public health targets 132 . For TT, active casefinding, including through mass population screening, is a recognized strategy for link ing individuals who have TT with surgical services, thereby achieving agreed public health targets 133 . Various screening approaches are used, including centralized villagelevel screening conducted by eye workers and doortodoor screening by TT case finders. The limited data available suggest that the sensitivity and specific ity of TT case finders are highly variable [134][135][136] . Properly training 134 and equipping 137 case finders is likely to be important.

Population-level assessment.
A key concept in trachoma epidemiology is 'elimination as a public health problem' , which is defined only for countries as a whole. It has three criteria: first, a prevalence of TT 'unknown to the health system' (that is, excluding individuals with postsurgical TT, individuals with TT who have refused surgery for it, and individuals with TT who have a surgi cal date set in the future) of <0.2% in those aged 15 years or more, in each formerly endemic district; second, a prevalence of TF of <5% in those aged 1-9 years, in each formerly endemic district; and, third, written evi dence that the health system can identify and manage incident TT cases 138 .
Although validation of elimination as a public health problem can only be undertaken for a country as a whole, districtlevel data are needed to substantiate the claim that the first two criteria have been satisfied. Here, the term 'district' means the normal administrative unit for healthcare management, generally a population unit of 100,000-250,000 people 138 . To avoid confusion with local administrative units designated as districts that may have populations outside the 100,000-250,000 range, trachoma programmes often use the term 'evalu ation units' instead. WHO has published recommenda tions on estimating the prevalence of TT and TF at the level of evaluation units 32,139 . Although rigorous surveys undertaken according to those recommendations were determined to cost a median of US$8,298 (in 2017) per evaluation unit 140 , the alternative approaches of simply continuing interventions without checking that they continue to be justified, or not eliminating trachoma at all, are likely to be considerably more expensive 10,141 .
Since 2012, the Global Trachoma Mapping Project (GTMP) 142 and its successor, Tropical Data 143 (TaBle 2), have together supported health ministries in 50 countries to conduct populationbased trachoma prevalence sur veys consistent with WHO recommendations, generat ing globally comparable, epidemiologically robust data 125 . Data from all GTMP and Tropical Data surveys are owned by the relevant health ministries, with automated linkages made available to display evaluation unitlevel prevalence categories on the openaccess Trachoma Atlas and pro vide the most current information to decisionmakers at all levels 144 .
Tests for Ct infection. Sensitive molecular approaches for pathogen detection, including qualitative and quantitative nucleic acid amplificationbased tests, are available 110,145 . In trachoma, they are generally used for research purposes and not for programmatic decisionmaking. However, they might find more wide spread use in the future as TF prevalence is not a perfect marker in this context, for several reasons 146 .
First, the sign TF is reasonably specific but not particularly sensitive for conjunctival Ct infection. In crosssectional surveys, a relatively high propor tion of individuals who are positive in nucleic acid amplificationbased tests for conjunctival Ct do not have TF 147 . The prevalence of Ct tends to be higher in those with TI than in those with TF, and higher in those with TF than in those with less florid signs of trachoma 54,83,104 . Some of this discrepancy is due to variation between the natural history of infection and the natural history of disease in the individual 3,110 . In particular, Ct infection is present for days to weeks before TF or TI develops, and is cleared days to weeks before TF or TI resolves.
Second, the specificity of TF for conjunctival Ct infec tion varies between contexts. In trachomaendemic pop ulations, less than half of those with TF are PCRpositive for Ct 147 . Again, some of this discrepancy will be due to the natural history of infection versus that of disease in the indvidual 110 , including the effect of frequent rein fection in accelerating Ct clearance after each infection episode 3 . In addition, other processes also cause folli cular conjunctivitis. Although in antibiotic MDAnaive  147 . In some countries in the Western Pacific, even before antibiotic MDA, children have moderately high prevalences of TF despite the extremely low prevalence of or absent conjunctival Ct infection [148][149][150][151][152] . Third, falling TF prevalences globally make the appropriate training of graders both more difficult and more important. Highquality photographs and central ized photograph grading 153,154 might help but would not solve the problem of the mismatch between the natural histories of disease and infection.
An appropriately designed 155 diagnostic test for current conjunctival Ct infection, deployed to help estimate infec tion at the evaluation unit level, could avoid misclassific ation from all three of these issues. Efforts have been made to develop rapid pointofcare diagnostics for the detec tion of current conjunctival Ct infection [156][157][158] . However, the principal programmatic use case is to determine whether or not interventions are indicated at the popu lation level; therefore, highquality tests for infection do not need to be performed next to the individuals recruited to participate in a survey. Alternatively, although open to misinterpretation in the assessment of a single individual, antiCt antibody data, based in particular on the presence of antibodies to the Ct antigen Pgp3, can be used to gen erate age-seroprevalence curves and seroconversion rates that are likely to be informative at the population level. This holds considerable promise for programmes 159 .

Prevention
Ct vaccine research is an important topic for basic and translational research 160 . From 2016 to 2017, a phase I, firstinhuman, randomized, placebocontrolled trial was undertaken of recombinant Ct protein CTH522, adjuvanted with either CAF01 liposomes or alumin ium hydroxide, as a potential vaccine against urogenital Ct infection. Both preparations seemed to be safe and immunogenic, with CTH522:CAF01 showing particu lar promise 161 . Currently, however, no vaccine is com mercially available that protects against conjunctival or urogenital Ct infection. Therefore, preventing blindness from trachoma relies on interventions intended to limit Ct transmission, treat Ct infection and reshape eyelids with TT so that eyelashes no longer touch the eyeball. These interventions are designed to achieve primary, secondary and tertiary prevention of trachomatous visual impairment.

Management
Interventions against trachoma are grouped together as the SAFE strategy: surgery (S) for TT, antibiotics (A) to clear infection, and facial (F) cleanliness and environ mental (E) improvement to reduce transmission 162 . The components of the SAFE strategy are delivered at different scales: surgery is offered to individuals, while anti biotics, facial cleanliness and environmental improvement are generally offered to whole evaluation units of 100,000-250,000 people. However, 'surgery' conceptualized as a public healthlevel intervention entails measures, sometimes including painstaking housetohouse case searches, to reduce the evaluation unitlevel prevalence of TT to below the threshold for elimination as a public health problem. Equally, it is crit ical that, while delivering the A, F and E components, the principles of nonmaleficence and autonomy for every individual are respected; as Addiss has written so power fully for another global health programme, we must continue to "see both the faces and the numbers" 163 of the people that we serve.

Surgery
The S of SAFE comes first not just because it makes a good acronym. Individuals with TT are at continuous risk of progressive visual impairment and are often in considerable pain 9 . Management to interrupt contact between eyelashes and the eyeball is urgent. Many different surgical procedures have been used to correct TT, with varying success 164 . The two proce dures recommended by WHO 165 are the bilamellar tarsal rotation procedure (BLTR) and the posterior lamellar tarsal rotation procedure (PLTR or modified Trabut). Both involve an incision across the upper eyelid, par allel to the eyelid margin, to release the trachomatous scarringinduced tension in the tarsus that draws eye lashes inwards, followed by placement of sutures to rotate the eyelashes outwards towards their normal anatomical position. Longterm surgical success is principally evaluated by assessing the cumulative inci dence of postoperative TT (PTT), which is caused by either immediate surgical failure or longerterm disease progression resulting in eyelashes touching the eyeball again. Other adverse surgical outcomes include pyo genic granulomata and eyelid contour abnormalities 166 . A pyogenic granuloma is a small, round, highly vascu larized growth on the conjunctival surface of the eyelid that develops as a result of injury or to protect against a foreign body such as a suture fragment left in the eyelid. Eyelid contour abnormalities are defined as any devi ation >1 mm from the normal curvature of the eyelid. These can range in severity from a slight change in the contour to substantial deviation in the form of a gap between the eye and eyelid, which can often lead to the development of a pyogenic granuloma.
PTT rates vary by procedure, geographical loca tion and type of study, with a cumulative incidence at 1 year after surgery ranging from below 10% to above 40% 167 . Surgeon skill level is an important predictor of outcome 168 but nonphysician surgeons can be as suc cessful as ophthalmologists 169 . A headtohead trial showed that PLTR was superior to BLTR in reducing the cumulative incidence of PTT, both in the short term and 3 years after surgery 170,171 . This trial included surgeons originally trained in PLTR who then converted to deliver BLTR, which could have influenced outcomes. Work to investigate whether BLTRexperienced surgeons should convert to PLTR is ongoing 172 . Currently, WHO recom mends that new surgeons should be trained in PLTR 173 , but existing surgeons trained in BLTR may continue to employ that procedure. Use of absorbable sutures and the TT clamp, which enables a single, safe, guided inci sion in the eyelid, have both been shown to reduce the incidence of pyogenic granuloma formation 174,175 . The TT clamp is also successful at reducing eyelid con tour abnormalities 175 . In settings where surgeons are highly skilled, a single dose of azithromycin reduces PTT incidence by about onethird 176 ; however, this effect is diminished in settings with increased rates of PTT 177 . Doxycycline does not reduce the incidence of PTT 178 . TT severity can range from a single eyelash touching the eye to the entire eyelid having cicatricial entropion with all eyelashes in contact with the eye. The number of trichiatic eyelashes present preoperatively and surgeon skill are the most consistent predictors of PTT, with more trichiatic eyelashes increasing the risk of developing PTT 168,179 . An analysis of multiple clinical trials also showed that peripheral trichiatic eyelashes

Box 2 | The diagnostic approach in a clinical encounter
In making a diagnosis for an individual patient, the clinician's primary responsibility is to that patient. Ancillary ethical or legal responsibilities may exist towards the clinician's own safety and well-being, the patient's contacts, the health system, or society as a whole, but bearing those other responsibilities in mind, a good clinician focuses all their diagnostic acumen to gain the greatest possible understanding of the patient's problems to help provide solutions. In formulating one or more diagnoses, the clinician, therefore, ideally takes all available information into account, including a complete history, general and specific examination findings, the epidemiological context, and the clinician's previous experience of the present and other patients.
For example, two preschool-age children are brought to the eye clinic of a district hospital, where a parent reports that each child has a history of a week or more of small amounts of eye discharge with crusting of the eyelashes and slightly swollen eyelids. The examining clinician notes the family's residence in an area known to be highly endemic for trachoma and their impoverished appearance. The clinician does not specifically look for trichiasis, as the probability of finding it is very low given the patients' ages and history. In both eyes of the first child, the clinician finds trachomatous inflammation -follicular and trachomatous inflammation -intense. The clinician diagnoses active trachoma and prescribes antibiotics to treat active trachoma. In each eye of the second child, there is inflammation, but neither the definition of trachomatous inflammation -follicular nor that of trachomatous inflammation -intense is met. The clinician prescribes antibiotics to treat active trachoma anyway, on the basis that the second child is in close contact with the first, and is highly likely to be developing active trachoma even if the current clinical appearance does not meet the criteria given in the World Health Organization (WHO) simplified system. were a risk factor for PTT regardless of the surgical pro cedure used 180 . The evidence on best approaches for the management of TT at first presentation is relatively clear but very little is known about how best to manage PTT. Individualized management by highly skilled surgeons is probably required 181 .
Since 2013, training programmes for TT surgery have evolved to maximize the skill levels of new sur geons. WHO now recommends that surgical simulation should be included in both new and refresher training programmes and that periodic monitoring of surgical quality after training is advisable 124 . Standard checklists are available to evaluate the skills required for each of the main surgical procedures 165 .
In trachomaendemic communities, surgery is typ ically provided free of charge, either through outreach campaigns or static services. Uptake of surgical services varies both within and between countries. Patients' understanding of what the surgery involves, distance to the location at which surgery is available and access to domestic help after surgery all affect uptake 182,183 . Surgical productivity is also affected by surgeons' access to surgical supplies and a supervisor 184 .
The removal of eyelashes with forceps is referred to as epilation, and this provides temporary pain relief. Crosssectional studies suggest that eyes with evi dence of previous epilation have a lower prevalence of corneal opacity 185,186 . Epilation is widely used in trachomaendemic communities by individuals who have TT or their caregivers (with caregiver involvement being particularly important for individuals with TT and impaired vision). A trial evaluating epilation versus surgery for eyelids with ≤5 trichiatic eyelashes showed that epilation was noninferior for vision outcomes but less successful at preventing eyelashes from touching the eye 187 . Epilation using highquality forceps and with regular followup is recommended where an individual with TT either declines surgery or has no immediate access to it 124 . Some countries have been reluctant to incorporate advice to epilate into their trachoma pro grammes based on concerns that it might further reduce surgical uptake. Evidence from Ethiopia suggests that programmatic support for epilation does not adversely affect the willingness of patients to consider future surgical management for TT 188 .
Although surgery is expected to primarily limit fur ther loss of vision owing to ongoing corneal opacifica tion, highquality surgery can actually improve visual acuity compared with the preoperative baseline 177,189 , presumably by reducing corneal oedema and tearing. TT surgery leads to marked decreases in pain, discharge and photophobia, improving both selfreported physical functioning and quality of life 190,191 , even in the absence of an improvement in visual acuity. Most PTT occurs within the first 12 months after surgery and subsequent incidence is low, suggesting that the longerterm prog nosis for patients who have good outcomes at 1 year is relatively positive 171,192 .

Antibiotics
Antibiotics are used to clear conjunctival Ct. Two alter native antibiotic regimens are recommended by WHO: 1% tetracycline eye ointment, instilled into the lower conjunctival sac of both eyes twice daily for 6 weeks, or a single oral dose of 20 mg azithromycin/kg body weight, to a maximum of 1 g (ref. 117 ). Evidence for the effect of oral azithromycin is stronger than that for topi cal tetracycline 193 . Topical tetracycline treatment also has several disadvantages that limit adherence, including a prolonged treatment course, which cannot be completely directly observed by healthcare workers. In addition, instillation into the conjunctival sac is difficult, and the ointment stings slightly on application and briefly blurs the vision after each administration. Oral therapy has the additional benefit of treating extraocular Ct infections that might otherwise be reservoirs for conjunctival recurrence 68 . Tetracycline eye ointment remains on the recommended list because it is very cheap, used in children aged <6 months of age, and is nearly univer sally available, whereas azithromycin is comparatively expensive if it is not donated.
The high intracellular concentration and long halflife of azithromycin are beneficial for singledose treatment of Ct, which can only replicate intracellularly. Initial trials showed that oral azithromycin was at least as effective as topical tetracycline for clearing infec tion in individuals 194,195 . Efficacy in a single dose and the excellent safety profile enabled consideration of its use in MDA. The Azithromycin in Control of Trachoma trial in Egypt, Gambia and the United Republic of Tanzania demonstrated that azithromycin MDA led to greater reductions in Ct prevalence than topical tetra cycline MDA 84 . This finding was supported by cohort studies demonstrating dramatic reductions in the preva lence and load of conjunctival Ct following azithromycin MDA 29,[196][197][198] .
The goal of the A component is to clear infection from as much of the community as possible, rather than merely treating individuals with clinically apparent active trachoma. Annual 199 MDA is employed, with a tar get of ≥80% coverage of the population at each treatment

Box 3 | The diagnostic approach in an epidemiological investigation
In a population-based survey or other epidemiological investigation, the clinician's primary responsibility to the individual survey examinee is to do no harm. Where diagnoses are made, the survey team has the responsibility of providing service or a pathway to service, but generating a comprehensive assessment of each examinee is generally not practical. A comprehensive assessment is also very difficult to fully standardize and is, therefore, inadmissible if trying to generate reproducible data. Thus, unlike the diagnostic process for an individual patient, information that is not in the survey protocol should not be used to weigh positive or negative findings.
In the morning of the day that the parent in the example provided in Box 2 had intended to take the children to the eye clinic, a team undertaking a population-based survey visits the household. The team is only surveying districts suspected or known to be endemic for trachoma and no weight is therefore given to the presumed epidemiological status of the population. Individuals resident in the household are asked to consent to examination; demographic data but no individual histories are recorded. The first child is examined: the findings are recorded as right eye negative for trachomatous trichiasis (TT -), positive for trachomatous inflammation -follicular (TF + ) and positive for trachomatous inflammation -intense (TI + ); left eye TT -, TF + , TI + . The second child is examined: the findings are right eye TT -, TF -, TI -; left eye TT -, TF -, TI -. Both children could still be given antibiotics for active trachoma by the survey team as both have eye problems that are likely to benefit from treatment, but it is critically important for the integrity of the data collection exercise and the most accurate estimate of population-level prevalence that the second child is not recorded as having TF, TI or active trachoma. round 117 . Azithromycin (specifically Zithromax, Pfizer, NY, USA) is donated to trachomaendemic countries through the International Trachoma Initiative 200 and is offered as tablets to individuals who are both ≥120 cm tall and aged ≥7 years. Azithromycin suspension is given to children aged ≥6 months and <7 years and individuals <120 cm tall; suspension can also be taken by those older and taller than these thresholds if they have difficulty swallowing tablets 201 . The 20 mg/kg body weight dose of azithromycin is approximated using a dosing pole, up to the adult dose of 1 g (refs 129,201,202 ). Children aged <6 months are offered tetracycline eye ointment.
Determination by trachoma programmes of the need for antibiotic MDA is based on evaluation unitlevel prevalence of TF in those aged 1-9 years. If prevalence is ≥30%, the evaluation unit is offered five annual rounds of antibiotics and then reevaluated with a repeat populationbased survey undertaken at least 6 months after the last round. If prevalence is 10-29.9%, the evaluation unit is offered three annual rounds, and if prevalence is 5.0-9.9%, a single round. Antibiotic MDA should be discontinued in evaluation units in which the TF prevalence, estimated through adequately powered prevalence surveys, is <5%.
The incidence of adverse events, such as local irri tation or skin reactions, with the use of tetracycline eye ointment is extremely low. Azithromycin is also very well tolerated. In MDA programmes, the overall incidence of reported adverse events following singledose azithro mycin treatment is <10% 203,204 . Most symptoms are minor and gastrointestinal, including abdominal pain, nausea, vomiting and diarrhoea 18,19 . In general, the incidence of adverse events increases with increasing age; in children aged <6 months participating in a study of azithromy cin MDA for prevention of mortality, the prevalence of adverse events did not significantly differ between chil dren receiving azithromycin or placebo 205 . Azithromycin MDA may have additional benefits for children, includ ing reducing allcause mortality (by 13.5% in the largest clusterrandomized trial 206 ) and the burden of some spe cific pathogens beyond Ct, including Streptococcus pneumoniae, Campylobacter spp., Treponema pallidum subsp. pertenue (the causative agent of yaws) and Plasmodium spp. [206][207][208][209][210][211][212][213][214][215][216] . Although observational studies have suggested a relationship between azithromycin use and risk of car diovascular death in adults in the USA 217 , a randomized study of MDA in Ethiopia found no increase in adult mortality in communities receiving azithromycin com pared with those receiving placebo, nor did it find a dif ference in mortality between individuals aged ≥30 years who took azithromycin and those who did not 218 .
To date, substantial macrolide resistance in Ct has not been documented following azithromycin MDA for trachoma [219][220][221] . Some evidence of selection for resistance in nontarget organisms exists 193,222 . In naso pharyngeal Streptococcus pneumoniae collected from populationbased samples of children, an increase in macrolide resistance has been found following azith romycin MDA, followed by a reduction in resistance when MDA was discontinued, removing selection pressure 223,224 . Selection for resistance in other potentially pathogenic organisms, including Escherichia coli and Staphylococcus aureus, has also been documented [224][225][226][227][228] .

Facial cleanliness and environmental improvement
The goal of the F and E components of the SAFE strategy is to reduce the transmission of conjunctival Ct. For F, the 'theory of change' holds that clean faces reduce the amount of infected eye and nose discharge available for transfer to fingers, fomites and flies, and reduce the attractiveness to female M. sorbens of both infected and uninfected eyes. Like the disappearance of trachoma from Europe and North America 12 , the elim ination of trachoma as a public health problem in the Islamic Republic of Iran has been attributed to improve ments in hygiene and environmental health 229 . Multiple crosssectional studies consistently show an association between a lack of facial cleanliness and the increased probability of having active trachoma [57][58][59] . However, these studies are limited both in temporality and by the potential for unmeasured confounding. The evidence base for the efficacy of interventions to improve facial cleanliness is not strong. A pairmatched, community randomized trial in the United Republic of Tanzania found no significant evidence that a hygiene promotion intervention designed to improve face wash ing among children reduced TF prevalence 230 . In this trial, the hygiene promotion intervention did reduce the prevalence of a marker of very severe active tra choma, the key driver to scarring, and having a clean face on multiple visits was protective against active trachoma. There is potential for reverse causality or unmeasured confounding between sustained facial cleanliness and active trachoma outcomes. A more recent clusterrandomized trial in Ethiopia found no evidence of an effect of hygiene promotion on the prev alence of conjunctival Ct infection in young children 231 . Interventions promoting facial cleanliness may be limited by difficulties in achieving sustained behavi our change, which is unlikely to be achieved using the approaches employed by most contemporary trachoma programmes 232,233 . Because evidence of efficacy for any particular intervention is weak, it is difficult to offer pro grammes an evidencebased recommendation to alter their approach; instead, it is only possible to observe that current approaches are not in accordance with established theory on behaviour change 234 .
Environmental improvement interventions attempt to increase access to water and the means to safely dis pose of human faeces. Theoretically, increased water access could improve personal hygiene, including face and hand cleanliness, and thereby decrease transmission, while improved sanitation could reduce the abundance of M. sorbens, which lays eggs on human faeces left exposed on soil 70 . The water used for personal hygiene does not necessarily have to be potable: seawater, for example, would (at least in theory) also be effective. Unfortunately, community randomized trials in Ethiopia and the Gambia have not provided evidence that the provision of latrines reduces the prevalence of active tra choma or conjunctival Ct infection [235][236][237] . Similarly, evi dence for the effectiveness of improving access to water on trachoma markers is inconclusive, lacking support from community randomized trials 231,237 .
Although a firm evidence base for the F and E com ponents of SAFE is lacking 237,238 , access to water and san itation is thought to be important for the management of many other neglected tropical diseases 239 as well as being a fundamental human right. A rightsbased approach mandates us to continue to promote the F and E compo nents of SAFE while simultaneously encouraging further research on specific interventions designed to reduce transmission of conjunctival Ct.

Quality of life
Active trachoma does not impair vision. Individuals with active trachoma tend to report mild eye irritation, per haps accompanied by a small amount of eye discharge that may cause eyelash crusting. Because these symp toms are common in trachomaendemic populations and because the peak prevalence of active trachoma is in young children, these symptoms often go unremarked 110 .
Conjunctival scarring causes progressive drying of the eyes 240,241 . This is associated with pervasive eye discomfort and fluctuating visual disturbances that can affect daily activities. Paradoxically, scarring of the lacrimal canaliculi can also lead to epiphora (apparent excessive watering of the eyes), as the normal drainage of the conjunctival sac is impaired and tear fluid escapes onto the cheeks 242 . Either dry eye or epiphora can have a considerable negative impact on quality of life 243,244 .
TT is painful to the point of being debilitating, even without objective loss of visual acuity. It has profoundly negative effects on quality of life, psychological health, productivity and social standing 9,112,113 that persist for as long as TT is present. For example, when assessed using WHOrecommended instruments, individuals with TT in Amhara Region, Ethiopia, had a substan tially lower overall quality of life (mean 34.5 versus 64.6; P < 0.0001) and health satisfaction (mean 38.2 versus 71.7; P < 0.0001) compared with control individuals matched for age, gender and location. These associations are present even in subgroups of patients and control individuals with normal visual acuity 112 . Surgery for TT improves the quality of life regardless of whether it improves vision 9,113,191 .

Progress against trachoma to date
From 2002 to 2021, the estimated number of people with trichiasis worldwide fell from 7.6 million to 1.8 million, a decrease of 77%. In the same period, the estimated number of people living in areas that warranted treat ment with the A, F and E components of the SAFE strategy for trachoma elimination purposes fell from 1.5 billion people to 136 million, a decrease of 91% 17 5). Despite this undoubtedly encouraging progress, it is important to acknowledge that the December 2020 deadline for the global elimination of trachoma as a pub lic health problem, agreed in 1996 (ref. 247 ), was missed; the target date has now been reset to 2030 (ref. 1 ). Of note, progress has been heterogeneous and elimination has proven difficult to achieve in some areas. This is further discussed below.

Future progress against trachoma
Elimination as a public health problem is a considerably more modest public health goal than eradication, which requires the permanent reduction to zero of the world wide incidence of an infection 248 . At an international meeting in 2019, most of the trachoma stakeholders who were present and surveyed believed trachoma could actually be eradicated at some future date 249 . Others believe that setting such a goal would not be in the best overall interest of global health, since the opportunity cost might be considerably greater than the immediate public health benefit 250 . This issue (like all other issues related to future progress against trachoma) will benefit from continuing input from all relevant stakeholders, particularly national programme managers. If eradiation was to be targeted, improvements in surveillance sys tems, increased funding, enhanced community engage ment and the absence of emergent antibiotic resistance (or development of an infectionblocking vaccine) may be prerequisites 249 .
An effective Ct vaccine would certainly change the landscape for trachoma. Although 100 years of work on chlamydial vaccines have not yet generated an effec tive product, encouraging data have emerged from a phase I trial of a recombinant protein subunit vaccine targeting the major outer membrane protein of Ct 161 . The lack of established immunological correlates of protection 161 and the relative difficulty of determining diseaserelated outcomes in the urogenital tract suggest that future partnership of vaccine developers with tra choma programmes may be rewarding: the protection offered by an effective vaccine will be easier to observe in the conjunctivae.
In addition to the further efforts to develop and make available an efficacious future vaccine, weaknesses of current programmes need attention, as several types of measurement error put further progress at risk.
The imperfect predictive power of TF prevalence as a marker for Ct infection at the population level ham pers progress; evaluation of indicators of conjunctival Ct transmission intensity other than TF prevalence is  44 countries were known to require interventions to eliminate trachoma as a public health problem, 7 required investigation to determine whether interventions are required, 3 claimed to have eliminated trachoma as a public health problem, and 12 had been validated as having eliminated trachoma as a public health problem. The boundaries and names shown and the designations used on this map do not imply the expression of any opinion whatsoever on the part of the authors or the institutions with which they are affiliated concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Adapted with permission from ref. 291 , World Health Organization (WHO).
an area of growing interest. Serological indicators may have an application in monitoring for recrudescence after antibiotic MDA has been discontinued, including in postvalidation surveillance [251][252][253] . Programmes need guidance on optimal strategies to detect possible recru descence but evidence on which to base this guidance is currently limited. Serosurveys would be one possible approach. However, currently assayed antiCt antibodies are not specific for trachoma Ct strains; thus, sero surveillance approaches using these antibodies would need to be driven primarily by data from children who have cleared maternal antibodies and not yet reached the age of sexual debut and would need to accept the contribution of inclusion conjunctivitis to population level seropositivity. Identification of an antiCt anti body specific to trachoma strains would be a major step forward in the utility of serosurveillance.
Another measurement issue arises from the difficulty of estimating the prevalence of a rare condition. Despite assiduous attention to epidemiological principles in sur vey design and implementation of quality control and quality assurance tools 126 , prevalence estimates gener ated by surveys remain estimates, subject to error. This becomes particularly important when public health decisions are made against fine margins as is inevitably the case when measuring progress against the elimina tion prevalence threshold (0.2%) for TT. The use of a geospatial rather than traditional frequentist statistical approach enables the harnessing of spatial correlation in prevalence data and may therefore improve the accuracy of TT prevalence estimates 254 .
In addition, important questions about each com ponent of the SAFE strategy remain to be answered. For the S component, data on the best approaches for managing PTT are needed as, even in the best hands, at least 10% of individuals who receive surgery for TT will commonly develop PTT 167 . Based on available global data, at least 180,000 individuals who had TT in June 2021 (ref. 17 ) will need further surgical management after their primary operations. In most settings, patients with PTT are currently managed with the same procedure that they initially received to manage their TT. Two surgical approaches for correcting PTT are currently being inves tigated in a clinical trial 255 , PLTR and the BevelRotation Advancement Procedure (BRAP) 256 . PLTR involves a partialthickness incision followed by rotation of the distal fragment to return eyelashes to their normal ana tomical position. In doing so, the distal and proximal tarsal fragments are overlapped, creating a thicker tar sus, which may be more likely to result in an eyelid con tour abnormality. BRAP is a new procedure designed to overcome this issue by creating a bevelled incision of the tarsus, dissecting between the anterior and posterior lamellae and removing scar tissue. This approach enables the marginal rotation to be combined with a posterior lamellar advancement and a reduction in eyelid thick ness. Additionally, some countries are beginning to rec ommend epilation as a definitive management strategy for PTT in which just one or two eyelashes touch the sclera temporal or nasal to the cornea 257 .
For countries to clear their TT surgery backlogs, efficient casefinding strategies are needed. The use of machine learning to develop image recognition software for TT identification is currently under way, which has the potential to improve broadscale screen ing for TT. Similar programmes are also being devel oped for the identification of active trachoma 258 . As TT becomes less common, strategies for ensuring access to highquality surgery for all who need it will continue to be important. This will require surgeons to maintain their skills through regular practise on simulators, and potentially other creative solutions for delivering inte grated, peoplecentred services 259 to a progressively more scattered population in need.
For the A component, research priorities depend on local epidemiology. In areas where Ct transmission persists despite years of implementation of A, F and E, such as some areas of Ethiopia 260 , new strategies (within the A, F or E components) for reducing community transmission of conjunctival Ct are needed 18 . In these areas, multiple rounds of annual azithromycin MDA may have produced a new equilibrium wherein trans mission continues but at lower levels than at baseline. Although increased antibiotic pressure, such as through more frequent MDA or a higher target antibiotic cover age, could hypothetically reduce infection prevalence, community randomized trials in Ethiopia and Niger have not consistently found empirical support for this theory [261][262][263][264][265] . However, in one trial in Ethiopia, quarterly treatment offered only to those aged 1-10 years achieved lower conjunctival Ct prevalence than annual MDA to people of all ages 266 , and evidence of a herd protection effect was noted with reduced Ct prevalence in untreated older children and adults 267 . This is likely to be a demon stration of a 'core group' effect in local transmission dynamics, in which suppression of transmission from and between children prevents infection from sustain ing an equilibrium level in the population as a whole 268 . Further studies are needed.
Work is also required to determine how to identify where (or even whether) antibiotic MDA is needed in evaluation units in which the TF prevalence is around the 5% elimination threshold. In communities in the United Republic of Tanzania with a baseline TF prev alence of 5.0-9.9% randomized to a single round of antibiotic MDA or no treatment, no difference in TF prevalence after 12 months was observed between study arms 269 . Longerterm study of the need for antibiotics in evaluation units where trachoma is disappearing or where the prevalence of TF seems to be rising again after antibiotic MDA discontinuation would provide greater insight into where and how antibiotics should be used for trachoma 270 . Tests for Ct infection need further exploration as potential guides for decisionmaking 271 .
For the F and E components, a better understand ing of Ct transmission and reliable markers of sus tained facial cleanliness would help in the testing of candidate interventions and assessment of programme effectiveness [272][273][274] . A community randomized trial of combined facial cleanliness and environmental improvement interventions did not prevent recru descence in conjunctival Ct infection prevalence after antibiotic MDA was discontinued in a trachoma hyper endemic population 231,275 , but it is important to note that coverage of interventions fell short of the 80-90% that models indicate might be a threshold for effectiveness 65 . Randomized controlled trials of WASH interventions are challenging to undertake, both logistically and in terms of achieving the transformative change in WASH access and use that is probably important 276 . However, such trials are essential to evaluate how best to deliver the F and E components to achieve sustained progress against trachoma. When washing the face, the use of soap may be important 277 .
The SAFE strategy is a combination of measures intended to prevent visual impairment from trachoma at multiple levels. Prevention of iatrogenic harm is also important and can be achieved through maintaining highquality surgery and exercising good antibiotic stew ardship. Surgical quality should be maximized through improvements in training, support, supervision and audit 184,278,279 of surgeons as well as consideration of who to operate on 188 . Good antibiotic stewardship involves avoidance of unnecessary antibiotic distribution 280 . Minimization of pressure driving the development of antimicrobial resistance 222 is growing in impor tance. Developing better tools for characterizing the populationlevel need for antiCt antibiotics for trachoma elimination purposes should help. Greater attention to visual rehabilitation and disability inclusion for those whose sight has already been affected by trachoma 281 also needs to be built into programmes. The effects of COVID19related interruptions to communitybased work on trachoma have yet to be empirically determined; models suggest that these may be considerable 282,283 .

Beyond trachoma
In January 2021, WHO published the road map for neglected tropical diseases 2021-2030 (ref. 1 ). A neglected tropical disease research and development blueprint, to be published as a companion document, is being pre pared. Research needs for trachoma, including some of the questions identified above, will form part of this. A common issue for all neglected tropical diseases will be how best to integrate what have often been relatively diseasespecific programmes to achieve maximum effi ciency, and how to further combine those efforts into a wholeofhealthsystem approach in collaboration with other sectors [284][285][286][287] . It is our hope that further characteriz ing the research needs of trachoma in this way, and then fulfilling those needs, will lead to a 2030 world in which trachoma is no longer a public health problem.
Published online xx xx xxxx