Original Article

The diameter of the nasolacrimal canal measuredby computed tomography: gender and racialdifferencesAustin McCormick FRCOphth and Brian Sloan FRANZCODepartment of Ophthalmology, Auckland Eye Department, Greenlane Clinical Centre, Greenlane, Auckland, New Zealand

ABSTRACT

Background: The incidence of dacryocystorhinostomysurgery among Pacific People is greater than wouldbe expected given their proportion in the population.Some have suggested that racial and sex differencesin facial skull dimensions produce narrower nasolac-rimal canals and therefore differences in the inci-dence of primary acquired nasolacrimal ductobstruction (PANDO). We measured the minimumdiameter of the canal in those not known to havePANDO.

Method: A retrospective review of the digital com-puted tomography (CT) database was performed.Minimum diameter of the nasolacrimal canal onaxial cuts of a sinus series CT scan was measured. Sexand racial origin were recorded. All series on thedatabase were included. Absence of axial images andpathology distorting the canal excluded a patient.This was carried out in the Department of Radiologyand Ophthalmology, Greenlane Medical Centre,Auckland.

Results: A total of 178 CT scans were included. Menhad a mean diameter of 3.9 mm (95% confidenceinterval [95%CI]: 3.8–4.1) versus women 3.6 mm(95%CI: 3.5–3.8) P = 0.01. Both Caucasian and NewZealand Maori had mean diameters of 3.7 mm(95%CI: 3.5–3.9) whereas Pacific People were4.1 mm (95%CI: 3.9–4.3) P = 0.01.

Conclusions: As in other studies women had nar-rower canals than men. Surprisingly we found no

difference between New Zealand Maori andCaucasian. Unexpected was the larger diameter inPacific People, as they have a higher incidence ofdacryocystorhinostomy surgery. PANDO is likely tobe of multifactorial aetiology and nasolacrimal canaldiameter may not be a significant factor. Ourdescribed method of measuring canal diameter by CTscan is comparable to a cadaver study.

Key words: canal, computed tomography, diameter,duct, nasolacrimal._ 357..361

INTRODUCTION

Primary acquired nasolacrimal duct obstruction(PANDO) is an idiopathic condition. It causes epi-phora and may require dacryocystorhinostomy(DCR). Wile the aetiology remains unknown, ana-tomical factors have been proposed. A recent audit ofDCR in Auckland showed a greater proportion ofDCR in women and Pacific People (Table 1).1 Otherstudies have suggested that racial and sex differencesin facial skull dimensions may explain narrowernasolacrimal canals and therefore differences in theincidence of nasolacrimal duct obstruction.2–5

Using the equation:

Flow resistance in a tube nL= ( )8 4πr

It can be shown that a 0.3 mm reduction in diam-eter of a tube increases resistance to flow of water by1.38 times. Although this equation is for a hollowtube containing water and clearly very different tothe in vivo situation of a canal lined with mucosa, itsuggests that small changes in diameter could be afactor in tear flow and thus canal obstruction.

� Correspondence: Mr Austin McCormick, Department of Ophthalmology, Aintree Hospitals NHS Trust, Walton Hospital, Rice Lane, Liverpool, L9 1AE,

UK. Email: austin.mccormick@aintree.nhs.uk

Received 22 September 2008; accepted 12 March 2009.

Clinical and Experimental Ophthalmology 2009; 37: 357–361 doi: 10.1111/j.1442-9071.2009.02042.x

© 2009 The AuthorsJournal compilation © 2009 Royal Australian and New Zealand College of Ophthalmologists

We designed a study to measure the nasolacrimalcanal diameter using computed tomography (CT) inthe major racial groups in Auckland.

We asked the following questions:

1 What is the diameter of the nasolacrimal canalmeasured by CT scan?

2 Are there gender differences in nasolacrimalcanal diameter?

3 Are there racial differences in nasolacrimal canaldiameter?

4 How does this method compare with otherstudies?

METHODS

We measured the minimum diameter of the nasolac-rimal canal in a population not known to havePANDO. Sinus CT scans were selected as the CTstudy protocol most likely to include an axial view atthe level of the nasolacrimal duct. Axial scans lieperpendicular to the canal allowing the diameter tobe measured most accurately.

All patients on the Auckland Agfa IMPAX digitalradiology database, accessed from Greenlane Radiol-ogy Department Auckland, who had a facial sinusCT scan from August 2002 to May 2006, wereincluded. Patients for whom there were no axialscans, pathology distorting canal anatomy and thoseless than 18 years of age were excluded. Their eth-nicity was recorded and we included: New ZealandEuropean; Other European; New Zealand Maori;Cook Island Maori; Fijian; Niuean; Samoan; Tongan.We excluded other racial groups as the numberswere small. Also recorded were age, gender anddiagnosis from the scan report. We measured theminimum diameter of the canal with the calliper toolof the IMPAX software. We scrolled up and down theimages including the nasolacrimal canal, choosingthe image that appeared to include the narrowestpart of the canal. The calliper software tool was oper-ated with a single mouse click at either end of thedistance to be measured. The start and end point of

the measure was the inner rim of the canal. Impor-tantly, the minimum diameter of the canal within asingle CT image was measured. The plane of the scanwas often not perpendicular, that is, oblique to thecanal and an elliptical cross-section was obtained. Inthis situation the minimum diameter in whateverdirection this lies is the actual diameter. To ensurethe minimum diameter had been measured, severalmeasurements were taken in different axes and theleast recorded. The frames above and below werealso measured to confirm the narrowest point of thecanal had been measured. The software automati-cally adjusted for different magnifications of scans.The callipers measured to 1/10th mm. Both right andleft were measured by a single investigator (Fig. 1).

Patients were written to and asked if they ‘hadever had tear duct surgery’ or if they had ‘a wateryeye causing tears to run down their cheek’ in anattempt to assess whether the population of peoplehaving sinus CT scans were in fact biased towardshaving nasolacrimal duct obstruction.

Ethical approval was gained from the Northern XRegional Ethics Committee, Auckland.

RESULTS

A total of 343 CT studies were reviewed. All weresinus CT scans. One hundred and seventy-eightstudies were included. The exclusions were: 152 – noaxial scans in the series; 9 – pathology distorting thecanal; 4 – aged under 18. The patient demographicsby age and racial description are shown in Table 2.Three main racial groups were analysed: Caucasian,New Zealand Maori and Pacific People. ‘PacificPeoples’ were those born and attributing their originto the island nations of the Pacific. Polynesian

Table 1. Frequency of DCR surgery by racial group in Auckland,compared with each group’s proportion in the population by theNew Zealand 1996 Census of Population and Dwellings

Patient subgroup DCR audit(%)

Population census(%)

Caucasian 49 62Pacific Peoples 28 11New Zealand Maori 12 12Male 32 49Female 68 51

DCR, dacryocystorhinostomy.

Figure 1. Axial computed tomography image from a sinusscan series. The software callipers have been used to measurethe minimum diameter of the nasolacrimal canal.

358 McCormick and Sloan

© 2009 The AuthorsJournal compilation © 2009 Royal Australian and New Zealand College of Ophthalmologists

includes New Zealand Maori and Pacific Peoples. Ascan be seen from Table 2, the smaller numbers ofsome of the racial groups made it impractical toanalyse them statistically, hence they were groupedtogether to form Pacific People. The mean diameterswere compared using a two-tailed Student’s t-test.

Table 3 summarizes the results of the nasolacrimalcanal diameter measurements. Caucasian were notsignificantly different from New Zealand Maori(P = 0.99), each with a mean diameter of 3.7 mm.Both Caucasian and New Zealand Maori weresmaller than Pacific People whose diameter was4.1 mm (P = 0.008 and P = 0.01 respectively). Menhad wider canals than women, 3.9 mm versus3.6 mm respectively (P = 0.01). No significant differ-ence in canal diameter was found between right andleft sides.

The CT scan reports were absent in 30%. Notsurprisingly sinusitis was the commonest finding in36%. Seventeen per cent had no abnormalityreported. Among the other findings were: nasalpolyps; frontal lobe abscess; hard palate cyst;dacryoadenitis; septal deviation; middle ear disease.

Sixty-eight per cent of questionnaire replies didnot report either epiphora or previous tear ductsurgery. Eight per cent confirmed that they hadundergone tear duct surgery and 32% reportedepiphora.

DISCUSSION

Surprisingly we found no difference between NewZealand Maori and Caucasian. Also unexpected wasthe larger diameter in Pacific People. As they have ahigher incidence of DCR surgery in the populationthan other racial groups, we were expecting a nar-rower canal. In our view, this suggests that nasolac-rimal canal diameter may not be a major aetiologicalfactor in PANDO. The same diameter canals forNew Zealand Maori and Caucasian may representmerging of the gene pools of these two racial groups.

Other studies using CT scans to measure nasolac-rimal canal diameter have shown larger gender andracial differences. A poster at 2005 ASOPRS scien-tific symposium described a similar study to oursin which the minimum diameter on an axial cutwas recorded. Women had an average diameter of3.78 mm and men 4.67. Caucasians had an averagediameter of 3.51 mm, Asians 4.44 mm and AfricanAmericans 4.67 mm.3 Another study in 1997 did notmeasure the minimum diameter but a standardanterior–posterior diameter which we believe hasled to over estimations when, as is common, a non-perpendicular scan image creates an elliptical cross-section of the canal.2

In 1969, RH Post measured the canal directly inprepared cadaveric skulls using wires of knowndiameter: Caucasian women (<50 years) 4.1 mm andmen (<50 years) 4.4 mm. He found that this differ-ence between men and women reduced with age to0.1 mm. He also found African American canals to be0.3 mm wider than Caucasian.4 One would expectthe mean diameters to be slightly different in the twomethods of measuring, as they are, but we areencouraged by the same difference between men andwomen in our method and the cadaver study. We feelthis confirms that CT measurement of the nasolacri-mal canal, as we have described, is an accurate, alter-native method of measuring this structure. In 1921 aCuban study also looked at racial differences in canaldiameter and detected similar differences betweenAfricans and Cubans as Post.5

It should be remembered that the nasolacrimalcanal describes the bony cavity whereas the duct isthe lumen with mucosal lining that resides withinand is bordered by the canal. If the canal diameteris not a significant aetiological factor in PANDOperhaps it is the duct that is pathological? A patienthistory should elicit known causes of secondarynasolacrimal obstruction: trauma, radiotherapy, sys-temic 5FU chemotherapy, topical eye drop use, infec-tion, inflammation, neoplasia.6

Chlamydia trachomatis is endemic in the Pacificislands and the WHO Western Pacific Region nowbears highest burden of Trachoma in the world.7 Anenvironmental agent such as this may explain thehigher incidence of PANDO in Pacific People;8

however, in a study which examined lacrimal sacbiopsies of 22 patients in Saudi Arabia with tra-choma, at the time of DCR, none were positive forChlamydia. They attributed this to inactive disease.

Linberg et al. propose idiopathic inflammation andsubsequent fibrosis,9 and a further study proposedmucosal ulceration leading to adherence andfibrosis.10

Hormonal differences between men and womenmay explain the greater incidence of this condition inwomen. A study, which measured tear outflow at

Table 2. Patient demographics by sex and racial description.Caucasian includes both New Zealand European and otherEuropean

Patient subgroup Male Female Total

Caucasian 46 53 99New Zealand Maori 17 15 32Pacific People 25 21 47Pacific People subgroups

Samoan 12 12 24Fijian 3 4 7Cook I. Maori 5 2 7Tongan 2 2 4Niuean 2 2 4Other 1 0 1

Nasolacrimal canal diameter by CT scan 359

© 2009 The AuthorsJournal compilation © 2009 Royal Australian and New Zealand College of Ophthalmologists

times during the menstrual cycle, found maximaloutflow during menses. It also suggests that sub-clinical hypothyroidism, whose incidence may be20% and is proportionately greater in women, maycause fluid retention and mucosal swelling.11 Zolliet al. found 43% of their women undergoing DCRhad previous hysterectomy or other cause forreduced oestrogen levels. They feel the nasolacrimalduct mucosa may become atrophic, in the same waythat the vaginal mucosa does in these patients. It isspeculated that this atrophy leads to loss of secre-tions and subsequent adhesions12

In 2000, capacitance vessels in the submucosa ofthe nasolacrimal duct were described in a study of 31cadaveric specimens. They form a plexus around thelacrimal sac and duct and drain to the cavernoustissue of the inferior turbinate. The histologyresembles a cavernous body in which the volume canbe altered dramatically. They propose that this struc-ture forms part of tear drainage regulation, that is,when vessels are engorged the tear outflow isreduced. This may be triggered by ocular surfacetrauma or disease, releasing inflammatory mediatorsand vasoactive substances. In other words, the epi-phora that occurs when a foreign body embeds in thecornea may be a combination of increased tear pro-duction and reduced tear outflow. Malfunction ofthis cavernous body may cause narrowing of the ductwhich could explain functional nasolacrimal ductobstruction (patent lacrimal syringing in a symptom-atic patient) or be a factor in the process leading toocclusion.13 In addition, it has been shown that adr-energic eye drops widen the nasolacrimal ductlumen and cholinergic eye drops reduce it. This auto-

nomic control or lack of it may be a factor in PANDO,perhaps by its action on the cavernous body.14

The incidence of epiphora is higher than expectedas is reported incidence of tear duct surgery. It ispossible that the replies are not accurate but even sothis demonstrates that we need to consider that thepopulation of those undergoing sinus CT scans maynot accurately reflect the general population. In otherwords the incidence of epiphora may be higher inthis group. This could bias the results. Our measure-ments may have underestimated the diameter ofnasolacrimal canals if this is a factor in duct obstruc-tion and our population has a greater than averageincidence of this problem.

A potential source of error in the study may bevariability of end-point in measurement, that is,where the measure was taken from and to. Great carewas taken to be consistent in placing the outer edgeof the on screen calliper arm up to but not over theinner edge of the bone. Bias was also possible.Although the ethnicity of the patient was notrecorded on the radiology database, the name wasand ethnicity may be inferred by surname.

Further investigation of this disease should lookat all possible factors including infection and inflam-mation of the soft tissue duct and its physiology. Amagnetic resonance imaging study may help to iden-tify nasolacrimal duct pathology. Histological exami-nation of the duct, particularly if at the time ofpossible active inflammation, may be instructive.

In conclusion, PANDO is likely to be of multifac-torial aetiology and nasolacrimal canal diameter maynot be a significant factor. The described method ofmeasuring canal diameter by CT scan is comparable

Table 3. Results of statistical analysis

All patients – gender differences All patients – comparing right vs. left

Male Female Male Female

N 87 91 t-test P value 0.82 0.94Mean 3.9 mm (3.8–4.1) 3.6 mm (3.5–3.8) Correlation 0.85 0.83t-test P value 0.01

Caucasian vs. Polynesian Caucasian vs. New Zealand Maori

Caucasian Polynesian Caucasian New Zealand Maori

N 99 79 N 99 32Mean 3.7 mm (3.5–3.9) 3.9 mm (3.8–4.1) Mean 3.7 mm (3.5–3.9) 3.7 mm (3.5–3.9)t-test P value 0.05 t-test P value 0.99

Caucasian vs. Pacific People New Zealand Maori vs. Pacific People

Caucasian Pacific People New Zealand Maori Pacific People

N 99 47 N 32 47Mean 3.7 mm (3.5–3.9) 4.1 mm (3.9–4.3) Mean 3.7 mm (3.5–3.9) 4.1 mm (3.9–4.3)t-test P value 0.008 t-test P value 0.01

95% confidence intervals are in brackets after the means. Three main groups were analysed: Caucasian, New Zealand Maori and PacificPeople. Polynesian = New Zealand Maori and Pacific People. Pacific People are those originating from the islands of the Pacific butexcluding New Zealand Maori.

360 McCormick and Sloan

© 2009 The AuthorsJournal compilation © 2009 Royal Australian and New Zealand College of Ophthalmologists

to a cadaver study and has shown women to havea narrower canal than men and Pacific People tohave wider canals than Caucasian or New ZealandMaori.

ACKNOWLEDGEMENT

We would like to thank The Radiology Department,Greenlane Clinical Centre, Auckland.

REFERENCES

1. Statistics New Zealand Te Tari Tatau. New Zealand1996 census of Population and dwellings. AccessedJanuary 2008. Available from: http://www.stats.govt.nz/census/1996-census-data/Default.htm

2. Groessl SA, Sires BS, Lemke BN. An anatomical basisfor primary acquired nasolacrimal duct obstruction.Arch Ophthalmol 1997; 115: 71–4.

3. Wladis EJ, Manelis JB, Carter SR et al. Gender andracial differences in lacrimal system anatomy. Posterpresentation: 2005 ASOPRS scientific symposium.

4. Post RH. Tear duct size differences of age, sex and race.Am J Phys Anthrop 30: 85–8.

5. Santos-Fernandez J. The measurements of the nasalcanal according to the race. Am J Ophthal 1921; 4: 32–7.

6. Imperia PS, Lazarus HM, Lass JH. Ocular complica-tions of systemic cancer chemotherapy. Surv Ophthalmol1989; 34: 209–30.

7. World Health Organization. Report of the 2nd globalscientific meeting on trachoma. Geneva, 25–27 August2003. Accessed January 2008. Available from: http://www.int/blindness

8. Rice CD, Kersten RC. Absence of chlamydia in tra-chomatous lacrimal sacs. Am J Ophthalmol 1988; 105:203–6

9. Linberg JV, McCormick SA. Primary acquired nasolac-rimal duct obstruction: a clinicopathologic report andbiopsy technique. Ophthalmology 1986; 93: 1055–62.

10. Hyde KJ, Berger ST. Epidemic keratoconjunctivits andlacrimal excretory system obstruction. Ophthalmology1988; 95: 1447–9.

11. Roussos J, Bouza A. Essai d’explication par des fac-tures hormonaux de la grande frequence d’apparitionde la dacryocystite chronique chez les femmes plutotque chez les homes. Bull Mem Soc Fr Opthalmol 1973;96–9.

12. Zolli CL, Shannon GM. Dacryocystorhinostomy: areview of 119 cases. Ophthalmic Surg 1982; 13: 905–10.

13. Paulsen FP, Thale AB, Hallmann UJ et al. The cavern-ous body of the human efferent tear ducts: function intear outflow mechanism. IOVS 2000; 41: 965–70.

14. Narioka J, Ohashi Y. Changes in lumen width of naso-lacrimal drainage system after adrenergic and cholin-ergic stimulation. Am J Ophthalmol 2006; 141: 689–98.

Nasolacrimal canal diameter by CT scan 361

© 2009 The AuthorsJournal compilation © 2009 Royal Australian and New Zealand College of Ophthalmologists