A Pubmed database search for relevant prevalence studies written in English and published as journal articles, between the years 1930 and 2018, was conducted. Individual and combined key words were used, including “convergence insufficiency,” “prevalence,”, “incidence”, “binocular vision anomalies”. Studies reporting prevalence of CI in patients with strabismus, late-onset CI, or who have undergone traumatic or acquired brain injuries, were excluded. Pertinent peer-reviewed articles from the primary search made on the basis of content and scope were then reviewed. Hand and manual searches were also performed for articles referenced in bibliographies that were not initially retrieved by the search, as well as in specific areas in which information was lacking. Meeting abstracts were excluded. Of the 260 publications retrieved, 33 were selected. Two of these were not pure prevalence studies, but were included because they stated the prevalence of CI in a cohort of participants. Another study reporting the prevalence of adult-onset CI, included participants aged 22 and above, and was therefore included as well.

Prevalence of CI reported in the literature (Table 1) differs among studies due to variations in outcome measures used to define CI, methodologies in obtaining outcome measures, age and gender distributions in samples, race, geographic locations, etc.. However, the relationships between these factors and prevalence of CI have not been investigated in depth6 though they may account for the wide variation of reported prevalence, ranging from 1.7% to 33%.

Diagnostic criteria vary among studies, from a single sign, such as a receded near point of convergence (NPC) ,4,7-10 to several signs.11-17 To demonstrate how this variability can lead to discrepancies amongst clinicians, clinical outcome measures of three representative patients extracted from the cohort of normal control participants from a previous study examining the prevalence of CI in whiplash associated disorder vs. controls,18 are tabulated in Table 2. Based on Duane's definition of CI19 (see Table 1), patient #1 would not be diagnosed with CI due to the break point, and patient #2 would not be diagnosed with CI due to the distance and near phoria. Based on Holland's20 four diagnostic criteria (see Table 3), all patients in Table 2 would be diagnosed as having CI.

Based on Scheiman and Wick's21 criteria (see Table 3), the cases in Table #2 only fulfill some of these required criteria and are missing several diagnostic outcome measures such as fixation disparity. Additionally, patients #1 and #2 do not have an NPC break point greater than 10 cm, and would therefore not be considered as diagnosed with CI. Based on the convergence insufficiency reading study (CIRS, Table 3)22 and the division into definite/ high suspect/ low suspect CI that has been adopted by others,12 patient #1 would be considered high suspect CI, while the other patients would be considered CI. Based on the Cochrane review of CI6 (Table 3) patients #2 and #3 have convergence reserves lower than 15Δ, whereas patient #1 has convergence reserves that are less than twice the near exophoria (XP) value. Note that if the requirement for convergence reserves was solely twice the amount of near heterophoria, patient #2 would not have been considered CI. In addition, the NPC value of patient #2 is not greater than the threshold value of 6 cm determined by the study.

Interestingly, the studies and textbooks tabulated in Table 3 did not include symptoms associated with CI as part of the diagnostic criteria. Based on the convergence insufficiency treatment trial (CITT, Table 3)17 diagnostic criteria, all three patients would have been considered symptomatic. Similarly, Kent and Steevev23 (Table 1) diagnosed CI based on asthenopia unrelated to refractive error alongside reduced convergence ranges. All three patients would have been diagnosed with CI based on this definition. Based on the diagnostic criteria listed in Elsayed and Abdou24 (Table 3), patient #2 with an NPC of 6 cm would not have been considered CI. Conversely, none of the patients reported diplopia as a symptom, and would not all have been considered CI according to the diagnostic criteria detailed by Ghadban et al.25 (Table 1). Even if diplopia had been reported by all three patients, only patient #1 has a near heterophoria greater than 10Δ, and would have been regarded as CI.

Symptoms can differentiate between symptomatic and asymptomatic CI.12 Patients with CISS scores ≥ 16 are classified as symptomatic CI, and those with lower scores are classified as asymptomatic.12,16 Based on this definition, all three patients can be considered symptomatic.

In an effort to address these discrepancies in diagnostic criteria, and to determine the optimal combination of clinical outcome measures for the diagnosis of CI in patients with a large near exophoria and moderate to severe symptoms, Cacho-Martínez and colleagues26 found that the combination of NPC recovery more remote than 8 cm and binocular accommodative facility less than eight cycles per minute yielded a sensitivity of 0.77, specificity of 1.00, and negative likelihood ratio of 0.23. However, few studies (see Table 1 and 3) list the binocular accommodative facility test as a diagnostic outcome measure for CI, and this information is missing for the patients in Table 2.

Thus, differing diagnostic criteria for CI can yield varying diagnoses in the same patients. Only one study assessed the diagnostic validity of specific clinical outcome measures by calculating their sensitivity and specificity, and examining the receiver operator characteristic curves.26,27 Researchers and clinicians should carefully assess diagnostic criteria stated in the study methods prior to drawing conclusions regarding treatment efficacy.

Even with very clear diagnostic criteria for identification and classification of CI, there is still considerable variability related to the specific clinical measurement techniques and cutoff criteria to be considered abnormal. Testing procedures should be standardized, because varying methods result in different norms and diagnostic values for NPC,28 heterophoria,29 AC/A30, fusional reserves,31 binocular and monocular accommodative facility,32 as detailed below.

The following section details different clinical assessments of CI and recommendations (in italics). The recommendations are based on the CITT study,17 with modifications for older age groups or in areas that are not addressed by CITT with appropriate references provided.

Heterophoria describes the deviation of the visual axes from perfect alignment (fixation) in the absence of fusion.18 Esophoria (EP), XP, and orthophoria describe axes joining before, behind, and directly on the object of regard, respectively.29 Heterophoria can be measured using the cover test, von Graefe, Maddox rod, Maddox rod with Thorington card for near,51 and Maddox Wing for near.52 The cover test includes two stages, unilateral cover/ uncover and alternating cover. In the unilateral part of the test, the examiner covers one eye and observes the movements of the fellow uncovered eye. If the fellow uncovered eye moves to obtain fixation, this movement compensates for the eye's deviated position. This is considered a strabismus. This process is repeated three times, before the contralateral eye is subsequently covered and the procedure is repeated. If both eyes do not exhibit a compensatory movement, there is no deviation, or strabismus and the patient is considered to have a heterophoria. The amount and the direction of the heterophoria are determined in the alternating cover test. In this test, the examiner alternately covers the eyes to break binocular fusion, while observing the eye that was occluded as soon as it is revealed and the cover moves to the fellow eye. If the eye that was occluded moves to obtain fixation, this means that the eye deviated when it was covered. The deviation can be measured with loose prisms or a prism bar18,29 as the amount of prism diopter (PD) necessary to neutralize the re-fixational movement.; the The cover test is considered objective because it does not rely on subjective patient responses like other clinical tests.18,29 It has been shown53,54 that the clinician's experience affects the results of the cover test and that the examiner endpoint criteria varies amongst individuals.55

Most (see Tables 1 and 3) but not all studies33 require distance and near heterophoria values for the diagnosis of CI. The CITT diagnosis criteria require a near heterophoria at least 4Δ more exophoric than the distance heterophoria, a condition that is met by all cases listed in Table 2. However, the method of measurement of heterophoria may not be consistent amongst studies. Differences in the measured values obtained with varying methods of testing may be insignificant clinically,29,56 however the methods differ in their variances. For example, the Maddox rod with Thorington card test was found to have better repeatability than the Maddox rod or von Graefe procedures for near heterophoria testing.51 Furthermore, varying outcomes are obtained under different conditions such as phoropter (“smooth”) vs. prism bar (“step”),56,57 cover duration,18 and varying values on the measurement scale (Maddox rod with Thorington card and Maddox Wing) .52,58

Some studies report that heterophoria tends to become exophoric with age.41,59 This tendency may be attributed to the reduction in accommodative amplitude with age.60 As detailed below, the accommodation and vergence systems are coupled in a reciprocal relationship that is quantified in the AC/A and CA/C ratios.61,62 Thus, when accommodation is reduced, the accommodative convergence is reduced, resulting is an increased XP. An increase in XP and incidence of CI with age was noted by Pickwell.63 A study of adult-onset CI in 118 participants over the age of 19 reported an increase of 7Δ of near XP over a 20-year time-period.25 Other studies report that the near heterophoria value varies with age in children.64,65 Conversely, constant values of adult distance heterophoria66 and near heterophoria values in children and adults have also been reported.65,67,68 Regardless of the effect of age on heterophoria value, the variability of the heterophoria measurement is dependent on the testing method and age of participants, demonstrating significant differences in presbyopic vs. non-presbyopic participants.

In an effort to maintain maximum objectivity, it is best to perform a prism-neutralized cover test to assess heterophoria at distance and near. Note that if only subjective tests are an option, a Maddox Rod at near with Thorington Card is the optimal choice.29 Near exophoria values greater than 6Δ 69–73 or that is at least 4Δ more exophoric than the distance value3,12-14,16,17,22,74-77 or that is decompensated are considered indicative of CI.

The AC/A ratio quantifies the amount of convergence obtained for every Diopter of accommodation activated.78 This can be calculated in the following manner. First, the difference between the near heterophoria and the distance heterophoria values is calculated. This value is multiplied by the near fixation distance, in meters. Then, the inter-pupillary distance, in centimeters is added. The resulting value, is the AC/A ratio.21 The gradient AC/A is the most common method, which is derived from the difference in the heterophoria value measured with and without spherical lenses, divided by the dioptric value of the spherical lenses.79 Scheiman and Wick recommend using either −1.00 or −2.00 DS spherical lenses,21 as implemented in some studies.80 Others studies used ±1.00 DS lenses,30 + 3.00 DS,81,82 or combinations of varying powers.83 At a fixed testing distance, tonic and proximal vergence are said to remain relatively constant, disparity vergence is minimal due to dissociation between the eyes, and spherical lenses placed in front of the patient alter the accommodative stimulus.30 The change in measured heterophoria divided by the spherical lens power denotes the AC/A ratio.21,41 The gradient AC/A ratio is a stimulus rather than a response AC/A, in which the amount of accommodation is assumed to correspond to the lens power placed in front of the eyes.84 The AC/A value has been identified as a distinguishing clinical sign for classifying BV anomalies21,42; a low AC/A value has been used as a diagnostic sign for CI by many studies.69-72,85 As discussed previously, heterophoria values may differ depending on testing method, especially repeated testing which is the foundation of AC/A measurement using the gradient technique.51,67,86,87 In a study comparing the effects of heterophoria measurement technique on gradient AC/A ratio, the lowest repeatability coefficient was obtained using the Maddox rod with Thorington card (MT) technique with both ±1.00 DS lenses, and the worst repeatability coefficient was obtained with the von Graefe technique with +1.00 DS lenses.30 A study comparing Howell Phoria Card (similar to MT) gradient AC/A to the response AC/A obtained using a Cannon autorefractor and −1.00 DS and −2.00 DS spherical lenses, reported poor correlation between the two measurements.80 A low correlation was also reported between the MT stimulus AC/A vs. response AC/A obtained using a Cannon autorefractor, though each method had good intra-test repeatability.88 The clinical cover test can be used to measure the gradient AC/A83 and at least one textbook recommends using a heterophoria measurement method that controls accommodation well.47 However, the use of the cover test has not been compared with the other subjective techniques.

Spherical lens powers can result in variable accommodative responses, yielding varying gradient stimulus AC/A ratios,89 though one study did not find differences in the response AC/A ratios with varying lens powers.90

Vergence ranges are measured using either loose ("step") or Risley rotary ("smooth") prisms.21,31,98 Horizontal vergence ranges can be examined at the base out (convergence) or base in (divergence) directions.41 A near target is held in front of the patient, imposing activation of a constant amount of accommodation. Prisms are placed in front of the patient's eyes and gradually increased until the patient reports blurring of the target. The blur point denotes the limits of fusional vergence with accommodation held constant Prisms are further increased until the patient utilizes the maximal accommodative vergence, and the target appears double. This is the diplopia point. The examiner reduces the amount of prisms until fusion is regained. This is the point of recovery. Some studies regard the convergence break point as a diagnostic sign for CI whereas others regard the convergence recovery point as a diagnostic sign.33

Studies show that the order of prism presentation affects the measurement and recommend the measurement of divergence prior to convergence.95,98,99

The type of test method (loose prisms- step vs. rotary prisms- smooth),31,98 target size,100 and examiner encouragement95 also affect the measured outcomes. This could well affect the diagnostic criteria for CI. The CITT17 protocol used loose prisms and a column of letters size 6/9 (20/30). Loose prism testing represents a more natural environment when compared with the smooth phoropter testing, which may be a reason for implementation of the step (loose prism) method.101Clinicians should adhere to one type of testing method for diagnosis and follow-up, as the testing methods are not interchangeable. The repeatability of both methods is better in the divergence as opposed to the convergence direction at both near and distance.31

Fusional reserves have been shown to relate to age and heterophoria value.102 Although only the recovery and not the break values appear to vary in adults,66 the convergence break values decrease significantly (∼8Δ) with age in myopic children.103

Another source of discrepancy for diagnosis of CI is the normative value of the fusional convergence reserves. For example, Porcar et al.69 did not define the values of the convergence ranges and included anyone with “low” reserves. However, others define reduced convergence reserves as either failure to meet Sheard's criterion (compensating fusional vergence should be twice the measured phoria)104 or values lower than a set amount which varies among studies. For example, Rouse et al.22 used break and recovery values lower than 12Δ/15Δ, Lara et al.70 and Hoseini-Yazdi et al.105 used values lower than 11Δ/14Δ/3Δ, Borsting et al.3 values lower than 7Δ/3Δ; Shin et al. ,71 Jang and Park,72 and Hashemi et al.14 used values lower than 12Δ/15Δ/4Δ; whereas Kent and Steeve,23 the CITT study group,17 Hussaindeen et al. ,15 Menjivar et al.16 and Stiebel-Kalish et al.77 used a break value lower than 15Δ. Davis et al.13 used 15Δ as the criteria for both break and blur values.

There is a reciprocal cross-coupling of the accommodation and vergence systems61,62 which is quantified in the AC/A and CA/C ratios, as described above. Due to the reduced convergence reserves in CI, patients may have been expected to demonstrate high accommodative amplitudes to overcome the lack of convergence by accommodative convergence.106 However, the studies listed in Tables 1 and 2 do not state accommodative amplitude as an expected diagnostic sign. Furthermore, in their textbooks, Scheiman and Wick state that the expected accommodative amplitude in CI patients is normal.20 One possible explanation could be the reduced AC/A ratio that characterizes CI69-72,85 providing little accommodative convergence gain for each Diopter of accommodative effort. The high accommodative effort required to overcome the lack of convergence may be inefficient for the visual system.

Co-morbidity of CI and accommodative insufficiency has been documented, especially in cases of severe CI. However, co-morbidity has been considered a separate entity and not true CI.107 In addition, CI with reduced amplitude of accommodation that is relieved with plus lenses is also not considered true CI, and has been called pseudoconvergence insufficiency.108

Accommodative facility is measured by asking the patient to report when the target is clear while the examiner alternates between ±2.00 DS lenses under binocular (BAF) or monocular (MAF) viewing conditions.21,109 Suppression checks are recommended for BAF testing.110 The number of alternations, or cycles per minute, is recorded. Normative values of 11 cycles per minute and 7 cycles per minute are typically used for BAF and MAF tests, respectively.43The BAF test assesses both the ability of the visual system to activate or release accommodation, as well as the ability to activate fusional vergence to compensate for the activation or release of accommodative convergence.21 The MAF test assesses only the accommodative system. Thus, if the patient fails the binocular but not the monocular test, the underlying problem is related to the vergence system. Several sources include failure of BAF with +2.00 Diopter lenses as a diagnostic sign for CI.15,21,70-73,85 If the patient fails both tests, the underlying problem is related to the accommodative system.21 Results of the accommodative facility tests improve with practice111 and are influenced by target type,112 target size,113 reaction time in alternating between lenses,112 and reaction time in patient responses.114 Thus, relying on BAF and MAF tests for diagnosing CI can be confounded by variations in factors related to testing methodology. Despite this, the combination of BAF with NPC recovery or break values was found to provide the best diagnostic validity for CI in patients with a large near exophoria that suffer from symptoms.26

Although the majority of the textbooks and studies agree there are symptoms associated with the CI, there is no clinical consensus regarding the symptoms. Von Graefe116 listed eyestrain, tension in or around the eyes, blurring and occasional diplopia during near work, closing one eye during reading for relieving ocular fatigue, and headaches. Duane19 described asthenopia, headaches, ocular pain, conjunctival irritation and spontaneous diplopia producing blurred vision at near. Hirsch117 stated that CI is accompanied by ocular discomfort or fatigue. The most common complaints of his sample of 48 university students with CI included ocular fatigue and general fatigue after sustained near work (38%), headaches after reading (25%), ocular aches/burning/stinging/tearing after near work (18%), and eyelid twitching (4%). However, Hirsch's diagnostic criteria for CI included low amplitudes of accommodation, which likely includes also accommodative insufficiency. As such, these symptoms cannot be attributed only to CI.

Davies2 described inability to concentrate while reading, distance photophobia, blurring of print, frontal or occipital headaches, nausea and fatigue alongside head retraction, sweating and distress upon request to converge. Kent and Steeve23 listed the prevalence of symptoms in their clinical population, the majority of which had more than one symptom. The most frequent symptom was headaches (60%), followed by blurring of print (49%), ocular fatigue (34%), occasional diplopia (21%), and other symptoms with unreported frequencies (e.g., nausea, eyelid burning, epiphora, sleepiness during reading, and loss of concentration).

Burian84 described pain and watering of the eyes, headaches, and asthenopia as a result of near work, and occasional diplopia. Mahto4 recognized symptomless and symptomatic CI subtypes, with symptomatic CI including eye strain, blurring of text, words running into each other during reading, ocular pain, headaches, and diplopia during near work.

Mohindra and Molinari118 pointed out that younger children are not expected to complain, and in these cases, clinicians should look for behaviors that may be associated with, or reflect effort to reduce diplopia, such as eyelid rubbing, head shaking, blinking, palpebral narrowing, or closing an eye. For older children and adults they recommend asking about complaints during sustained near work, including horizontal diplopia or running of words in a line.

Others describe symptoms associated with general binocular visual disorders, and not specifically CI. These include asthenopia, headache, blur, and diplopia.119 A questionnaire assessing the frequency, severity and association of asthenopia was suggested by Sheedy and Saladin119 in an effort to recognize clinical signs that provide the best indication of binocular visual problems, but was not included in their publication.

Porcar and Martinez-Palomera69 reported that the most common (8–20 percent) symptoms in 65 university students were asthenopia, headaches, photophobia, and blurred vision either at distance/near/or transition between distances. Less common symptoms (∼3%) included diplopia and poor concentration. No correlation analyses with specific or general binocular visual dysfunctions were performed.

Abdi et al.10 screened 216 school aged children (1st, 4th and 8th grade) for refractive errors, binocular visual functions, and subjective symptoms using a self-reported questionnaire. Questions included fatigue and poor comfort during reading and writing, double vision or loss of words during reading, blurred vision or trouble focusing at/near a computer, and headaches during reading or at the end of the school day. Although 23.1 percent reported asthenopic symptoms, these were not significantly correlated with binocular visual problems and were significantly correlated with low uncorrected visual acuity and myopia.10

Westman and Liinamaa120 retrospectively examined symptoms of 135 patients with CI aged 6–79 years before and after orthoptic treatment. Reported symptoms included difficulty performing near work (69%), headaches (63%), reading difficulty (46%), ocular fatigue (24%), diplopia (21.5%), ocular pain (15%), and watery eyes (3%). Orthoptic exercises eliminated symptoms in 60% of the children (≤18 years of age) and 52% of adults.

The CIRS group found that children with convergence and/or accommodative insufficiency often report symptoms of blurred or double vision, with probability of reporting symptoms increasing with the number of diagnostic clinical signs.3 They developed the CISS to analyze scope and severity of symptoms in order to diagnose and follow-up improvement with treatment. The questionnaire has been validated in children121 and adults 19–30 years.48 The questionnaire comprises 15 questions to which the patient must respond (never, infrequently, sometimes, fairly often, and always) to provide a total score. The cut off criteria for the CISS varies between children (≥16) and adults (≥21). It has also received wide application for BV disorders other than CI,122 and been shown to be a general symptoms questionnaire not specific to CI.123

One of the difficulties with self-reported symptoms is differentiating between symptoms related to BV anomalies vs. other causes. This was highlighted in a study examining the prevalence of CI and symptomatic CISS scores (≥21) in a sample of 171, 18–26 year old university students124; 41(25%) had symptomatic CISS scores, 11 had CI, but only six of those with CI also had symptomatic CISS scores. The authors concluded that symptomatic CISS scores are common in this population and not indicative of CI, basing their diagnosis of CI on a receded NPC > 8 cm from the bridge of the nose (not 6 cm from the forehead used by the CITT), concurrent with convergence reserves blur/break < 12Δ /15Δ. They excluded participants with large XP at near, without defining a large XP. As such, the authors' definition of CI differs from the definition of CI based on the CITT protocol, and their conclusions do not necessarily apply to the CITT definition of CI.

The facility of the vergence system is examined by rapidly switching between Base Out (BO) and Base In (BI) prisms that are placed in front of the patient's eyes. The prism powers are traditionally 8 Δ BI and 8 Δ BO, 5 Δ BI and 15 Δ BO, or 12 Δ BO and 3 Δ BI.125 The examiner counts how many times the prism base can be alternated (each alternation is a cycle) during the course of one minute, while the patient is able to fuse and see the text clearly. The combination of 3Δ BI and 12Δ BO for vergence facility testing has been shown to be most repeatable at near.125 Vergence facility examines the ability of the vergence to rapidly change over time,125 and those suffering from CI are expected to encounter difficulty in sustaining the BO direction.21