Incidence and Clinical Presentation of Moderate to Severe Graves' Orbitopathy in a Danish Population before and after Iodine Fortification of Salt

Abstract

Context:

Population-based data on the incidence and clinical presentation of moderate to severe Graves' orbitopathy (GO) are scarce, and virtually nothing is known on the effect of an iodization program on the incidence and presentation of GO.

Objective:

The objective of the study was to characterize incident moderate to severe GO in North Jutland County, Denmark, during the period 1992–2009, before and after the Danish salt iodization program.

Design and Patients:

The design of the study was a prospective register of patients with incident moderate to severe GO in a population during 8.9 million persons × years of observation.

Setting:

The study was conducted at a thyroid-eye clinic of university hospital.

Main Outcome Measures:

Clinical presentation and incidence before and after the year 2000 initiation of the mandatory Danish iodization of salt were measured. The incidence of GO was related to the incidence of Graves' hyperthyroidism (GH) in the same population.

Results:

The incidence rate of moderate to severe GO was 16.1/million per year (women: 26.7; men: 5.4), with no change associated with iodization of salt. The moderate to severe GO incidence was 4.9% of the incidence of GH. The incidence rate ratio between women and men with GO (4.9) was not different from the ratio in GH. Compared with GH, only a few patients (<2%) suffered from moderate and severe GO below the age of 40 yr, whereas GO was relatively common in age groups 40–60 yr (∼8%).

Conclusions:

Approximately 5% of the patients with Graves' disease develop moderate to severe GO, with a similar risk in women and men with Graves' disease. The risk of GO is much higher in patients aged 40–60 yr than in young patients with Graves' disease. Salt iodization was not associated with a change in the incidence of GO.

Graves' disease is a common autoimmune disorder that may present with various clinical manifestations. A key element in disease development is autoimmunity against the TSH receptor, although the mechanisms involved in disease outside the thyroid gland are only partly known (1, 2). The most common manifestation of Graves' disease is hyperthyroidism caused by TSH receptor-activating antibodies. Approximately 20% of patients suffering from newly developed Graves' hyperthyroidism have some degree of eye symptoms and signs (3) that are most often self-limiting and transient. However, a subgroup of patients develop moderate to severe orbital autoimmune damage that may impair eye function in several ways (1, 4) and that leads to a pronounced reduction in quality of life (5, 6).

Population-based data on the incidence and clinical presentation of moderate to severe Graves' orbitopathy (GO) are scarce (7), and virtually nothing is known on the effect of the iodization programs that are in effect in many countries (8) on the incidence and presentation of moderate to severe GO. One of the countries that has initiated an iodization program in recent years after being iodine deficient for many years is Denmark (9). In the year 2000, a mandatory program of iodization of salt for households and for commercial bread production was introduced in Denmark.

The aim of the present study was to characterize incident moderate to severe GO in North Jutland County, Denmark, during the period 1992–2009, and to compare incidences and presentation before and after the Danish salt iodization program.

The Danish iodization program is monitored by the Danish Investigation of Iodine Intake and Thyroid Diseases (DanThyr) program (9) that includes the collection of data on the incidence of overt thyroid dysfunction in two areas in Denmark. We used newly published data on the incidence of hyperthyroidism caused by Graves' disease in North Jutland County (10) to relate the incidence of moderate and severe GO to the incidence of Graves' hyperthyroidism.

Materials and Methods

Graves' orbitopathy patients studied

The collaborate Thyroid-Eye Clinic of Aalborg Hospital was established in 1990, and patients have been systematically investigated using the same protocol since January 1992. From January 1, 1992, to December 31, 2011, we investigated 203 new patients with active moderate to severe GO in the clinic.

In the present study, we aimed to include all patients with moderate to severe GO in a certain population. We therefore selected patients who lived in the primary catchment area of the Thyroid-Eye Clinic of Aalborg Hospital, which is North Jutland County. In this population there is only this Thyroid-Eye Clinic approved by the Danish National Board of Health (supreme professional health care authority in Denmark) to care for patients with moderate to severe GO, and throughout the study period, there has been no indication that any patient living in the county was treated outside the county without being seen in our clinic. Patients primarily seen at the Department of Endocrinology for thyroid function abnormality but who subsequently developed moderate to severe GO were included similar to other patients. All patients included were judged to have active GO.

Moreover, we aimed to study the incidence of moderate to severe GO before and after the Danish mandatory iodization program of salt that was in effect from the year 2001 and to compare the incidence of GO with that of Graves' hyperthyroidism in the same area. Therefore, we included patients who were investigated during the 9-yr period before iodization of salt (January 1, 1992 through December 31, 2000) and the 9-yr period after (January 1, 2001, through December 31, 2009). The periods studied and the median urinary iodine concentrations in the Danish population before and after the initiation of salt iodization are illustrated in Fig. 1.

Fig. 1.

Schedule of the Danish mandatory iodization of salt in relation to the period of inclusion of patients with moderate to severe GO. The patients were included during 9-yr periods before and after salt iodization. Salt iodization increased the median urinary iodine concentration in the Danish population, as indicated elsewhere (34).

By January 2007 Danish counties were replaced by regions as administrative units. However, GO patients admitted during the period 2007–2009 were included only if they lived within the previous county boundary. For calculations of incidences, we used population data from the Danish statistical bureau. By January 1, 1996, the county of North Jutland had 490,836 inhabitants, and by January 1, 2005, it was 495,068 inhabitants. For calculating average incidence during the entire 18-yr period, we used the population by January 2001, which was 494,833. Data on smoking habits in the Danish population (1996: 38%; 2005: 25% regular smokers in people ≥ 13 yr of age) were from Danish National Board of Health (www.sst.dk).

Characteristics of new patients with Graves' orbitopathy were systematically and prospectively studied during the entire period, as described in the European Consensus Report on GO (4), and if patients were classified as having moderate to severe GO, their data were included in our registry. Principles of classification were developed before 1992 and they were identical during the study period. Patients classified to suffer from active moderate to severe GO were considered for specific therapy using medical immunosuppression and orbital radiation, whereas patients with mild orbitopathy were not (4).

Characterization of the orbitopathy

A systematic investigation was performed by both an endocrinology and an ophthalmology specialist to characterize the disease and evaluate the possibility of other interfering diseases, and orbital CT or MR scans were performed. Over time, many classification systems for GO have been used (11–18). We had no intention of developing another classification system for general use, but prospectively recorded various components of the orbitopathy using somewhat simplified measures suitable for our daily clinical practice. The system used consisted of recording the following: 1) affection of visual acuity, 2) signs of inflammation, 3) eye lid abnormalities, 4) proptosis, and 5) eye motility and double vision.

Patients were classified as having sight threatening orbitopathy in case of subjective loss of vision on at least one eye in relation to development of GO and the finding of a visual acuity of 0.67 or less with no other cause being detected.

Signs of orbital inflammation were recorded using a 0–4 score (0: none; 1: mild redness and swelling; 2: moderate redness and swelling; 3: considerable redness and swelling with chemosis; 4: severe chemosis, swelling, and redness). Eyelid abnormalities were measured by ruler as millimeters of visible sclera below and above the central corneal limbus with the head and eye in neutral position, and as millimeters of closure deficiency (lagopthalmus). Eye proptosis was recorded using Hertel exophthalmometry.

Motility was recorded for each eye by drawing on a simple circular diagram split into quartiles, with the circle representing expected motility in each direction. Normal eye motility in various directions ranges between approximately 40° (up) and approximately 60° (down) depending on measurement method used (19, 20). In patients with active GO, measurement of eye motility is frequently hampered by the induction of pain from stretching of affected eye muscles. No system has been developed to record restriction of individual eye muscles independent of affection of other muscles (19). We asked the patient having the head in a fixed neutral position to look fixedly at a 1-cm object held in a 60-cm distance. The object was subsequently moved from neutral to maximal fixation up and down, and at the extremes to the right and to the left, and then from neutral to maximal right and left, and at the extremes up and down. This was first done binocularly to compare motility of the eyes and then for motility recording with coverage of right and then left eye. Judgment of motility was assisted by watching the position of the light reflex on the surface of the eye (18). Eye motility was recorded as normal (0% reduction) if judged to be normal by the investigator and if the patient experienced no abnormality during the testing of eye motility. A motility reduction was subsequently measured from the diagram by planimetry using a grid in which each crossing represented 5% of the expected field of motility. Similarly, the patient's subjective field of double vision was recorded on a circular diagram by drawing the binocular fields of single and double vision during eye movement for fixation, and the relative field of double vision was measured from the drawing by planimetry. All recording of motility and double vision used for the present study had been done by the same investigator (P.L.).

In the individual patient, the disease was normally characterized by a combination of manifestations of various severity. We recorded which manifestation was the dominant cause for the classification as moderate to severe GO in each of the 143 patients: 1) impaired vision (worst eye), 12.6%; 2) double vision affecting at least 40% of binocular field of single vision, 55.2%; 3) restricted eye movement: 40% or greater of uniocular field of fixation (worst eye), 13.3%; 4) proptosis 25 mm or greater (Hertel) (worst eye), 7.7%; and 5) combinations including severe inflammation (worst eye), 11.2%. Manifestations were evaluated in sequence as listed and only lowest number used. Distribution was not different before vs. after iodization of salt (P = 0.72, χ² test).

Hyperthyroidism caused by Graves' disease

The incidence of hyperthyroidism caused by Graves' disease was previously studied in two Danish population cohorts as part of the monitoring of the Danish iodization program (DanThyr) (10). This study included a subpopulation of North Jutland County (n = 311,102 by January 1, 1999), which was monitored during the period March 1, 1997, to December 31, 2000. Methods and results have been described in detail (10). In brief, subjects with thyroid function tests suggesting newly developed overt hyperthyroidism [low serum (s)-TSH combined with high s-T₄ and/or s-T₃] were identified from a computer-based monitoring of all results of thyroid function tests performed in the population, and the diagnosis was verified by contact to the physician requesting the tests (21). Subsequently we collected information on medical history, thyroid hormone receptor antibody measurement, and thyroid scintigraphy. Information was used to disprove or verify primary overt hyperthyroidism and to subclassify into nosological types of hyperthyroidism. The sex- and age-standardized (to the Danish population) incidence rate of hyperthyroidism caused by Graves' disease was 312 new cases per 1 million inhabitants per year, and the standardized incidence rate ratio between women and men was 4.5 (10).

Statistical analysis of data

Incidence rates (IR) and IR ratios as well as confidence intervals (CI) were calculated according to standard procedures (22, 23). We used SPSS (version 15.0; Chicago, IL) and Microsoft Excel 2003 (Redmond, WA) for calculations and analysis of data. Type of statistical test is indicated where used.

Data processing was approved by the Danish Data Protection Agency.

Results

Incidence of moderate to severe GO

Over the 18-yr period 1992–2009, the registration of new cases of active moderate to severe GO included 143 patients living in North Jutland County. This corresponds to an incidence rate of 16.1 per million per year. All patients had a Caucasian ancestry except for one male of Asian ancestry.

The number of female patients was 119 corresponding to an IR of 26.7 per million women per year and the number of men was 25 (IR 5.4 per million men per year) [IR ratio (women to men) 4.9].

Incidence rates were identical before (15.9 per million per year) and after (16.3 per million per year) iodization of salt. The absolute number of patients included before and after iodization of salt were 68 and 75, respectively.

Comparison with Graves' hyperthyroidism

The epidemiological study of hyperthyroidism in North Jutland (Denmark) included 393 new patients with Graves' hyperthyroidism (women, n = 321; men, n = 72) (10). The women/men ratio was similar in GO and Graves' hyperthyroidism (Pearson χ² test, P = 0.68).

Age distributions were different in patients with GO and Graves' hyperthyroidism as illustrated in Fig. 2. Compared with Graves' hyperthyroidism, only a few patients (<2%) suffered from moderate to severe GO below the age of 40 yr, whereas GO was relatively common in age groups 40–60 yr (∼8%).

Fig. 2.

Ratio between incidence rates of moderate to severe Graves' orbitopathy and of Graves' hyperthyroidism according to patient age. The horizontal stippled line indicates the incidence rate ratio (4.87 × 10⁻²) at all ages combined and the shaded area around the line is the 95% CI (4.02–5.90) for this value. The incidence rate ratio (95% CI) for patients younger than 40 yr was 1.16 (0.63–2.15), and for patients 40+ yr of age, it was 6.50 (5.27–8.01).

Overall, the incidence rate ratios (×10²) between GO and Graves' hyperthyroidism were 5.0 in women and 4.5 in men (women + men: 4.9).

Characteristics of patients with GO

Table 1 shows detailed characteristics of patients included before and after iodization of salt and for the entire period. No differences were observed in characteristics between the two study periods except for a moderate reduction in frequency of current smoking, from 84% during the first period to 62% during the second period. Based on data published by the Danish National Board of Health (www.sst.dk) on smoking habits in the Danish population and population data from the Danish statistical bureau (www.dst.dk), we tested whether the incidence rate of GO in smokers and nonsmokers had changed from the first to the second period but found no statistical significant differences (data not shown).

Table 1.

Patients with moderate to severe GO in North Jutland County before and after salt iodization

	1992–2000 Before I	2001–2009 After I	Before vs. after P	1992–2009 All patients
New cases, n	68	75	0.35a	143
Women/men, n	58/10	61/14	0.65b	129/24
Age (yr)c	50 (46–58)	56 (45–63)	0.16d	53 (46–62)
Weight (kg)c	69 (60–84)	70 (60–80)	0.78d	70 (60–82)
Hyper/eu/hypothyroid, n	61/4/3	64/6/5	0.73b	125/10/8
GO duration (months)c	7 (5–12)	7 (4–11)	0.28d	7 (5–11)
Thyroid dysfunction duration (months)c,e	12 (7–24)	13 (5–25)	0.84d	12 (6–24)
Current smoker, n (%)	56 (83.6)	47 (62.3)	0.003b	103 (72.0)
Lateralization score (0–3), nf	41/13/11/3	32/22/17/4	0.21b	73/35/28/7
Visus impairment, n (%)	10 (14.7)	8 (10.7)	0.62b	18 (12.6)
Inflammation score (0–4), n	1/11/25/23/8	3/12/17/24/19	0.17b	4/23/42/47/27
Sclera upper visible, n (%)	42 (61.8)	36 (48.0)	0.17b	78 (54.5)
Sclera lower visible, n (%)	25 (36.8)	21 (28.0)	0.42b	46 (32.2)
Closure deficiency, n (%)	22 (32.4)	22 (29.3)	0.72b	44 (30.8)
Double vision, n (%)g	51 (76.1)	55 (75.3)	0.92b	106 (75.7)
Double vision (% of field)c,h	70 (50–85)	65 (50–85)	0.98d	70 (50–85)
Motility reduction, n (%)	64 (94.1)	71 (94.7)	1.00b	135 (94.4)
Motility reduction (% of field)c,h	50 (30–65)	50 (30–60)	0.43d	50 (30–65)
Hertel worst eye (mm)c,i	23 (21–26)	23 (21–25)	0.40d	23 (21–25)

^aNormal test.

^bPearson-χ²/Fisher's exact test.

^cMedian (25–75%).

^dMann-Whitney U test.

^eOnly patients with thyroid dysfunction were included.

^fLateralization scores were: 0, no lateralization; 1, slight lateralization; 2, pronounced lateralization with only mild GO on one side; and 3, entirely one sided.

^gThree patients with monocular vision were excluded.

^hPatients with no affection were excluded.

ⁱHertel measurement was not possible in one patient.

The majority of GO patients (87%) also suffered from hyperthyroidism, whereas 7% were euthyroid and 6% were hypothyroid. In most patients with hyperthyroidism, this had developed before or simultaneously with the orbitopathy (Fig. 3). The orbitopathy was symmetric with no side difference in approximately half of the patients, whereas the other half had some degree of asymmetry, with 5% presenting entirely unilateral disease. Signs of orbital inflammation were seen in nearly all patients, although no such signs were observed in a minority of 3%. Approximately half of the remaining patients had slight to moderate inflammatory signs, whereas the other half had considerable and severe signs (score 3 and 4). Approximately 70% of patients had some degree of lid abnormalities, most often a visible sclera between the upper lid and the cornea. Nearly one of three had some closure deficiency. Restriction of external eye muscles with motility reduction and diplopia was very common. As shown in Fig. 4, 94% of the patients had some restriction in uniocular field of fixation recorded (Fig. 4, upper panel), whereas 76% of patients had some degree of double vision (Fig. 4, lower panel).

Fig. 3.

Temporal relationship between diagnosis of thyroid dysfunction [hyperthyroidism (Hyper) or hypothyroidism (Hypo)] indicated by the 0-arrow and debut of Graves' orbitopathy in 133 patients with moderate to severe Graves' orbitopathy and with thyroid dysfunction. Note the change in time scale of abscissa indicated by stippled lines.

Fig. 4.

Upper panel, Frequency of uniocular restriction in field of fixation (worst eye) (percentage of expected field of motility). Lower panel, Reduction in binocular field of single vision. For evaluation of double vision, three patients with monocular vision were excluded.

Figure 5 shows the distribution of Hertel measurements in men and women. Approximately 85% of the patients had values above the upper normal for Caucasians.

Fig. 5.

Distribution of Hertel exophthalmometer readings (worst eye) in women and men with moderate to severe GO. One patient was excluded for technical reasons. Hertel readings in women were 22.6 ± 3.2 mm (mean ± sd), and they were 24.2 ± 3.1 mm in men. Assuming an upper normal value of 19 mm in women and 21 mm in men (34), 82.2% of women and 87.5% of men had elevated values.

Discussion

Principal findings

This observational prospective incidence study performed in a Danish county found no time-dependent trend in the incidence of moderate to severe GO during the period 1992–2009. Specifically, neither incidence nor presentation of disease differed before and after the mandatory Danish iodine fortification of salt. The only difference observed was a moderate decrease in the frequency of smoking corresponding to the general trend in the Danish population. Still, smoking was very common in the patients, which is in accordance with a number of previous studies (1, 7, 24).

The incidence of moderate to severe GO was 16 cases/million per year, which was approximately 5% of the incidence of hyperthyroidism caused by Graves' disease, with a similar sex ratio (women to men ratio between 4 and 5) in the two manifestations of Graves' disease. Age distribution differed in GO and Graves' hyperthyroidism, with GO being less than 2% of Graves' hyperthyroidism at age 20–40 yr and 8% at age 40–60 yr. In the majority of patients who suffered from both GO and Graves' hyperthyroidism, the debut of GO came after the debut of Graves' hyperthyroidism. Nearly all patients with moderate to severe GO had clinical signs of external eye muscle restriction, and double vision was the most common cause for the moderate to severe classification.

Comparison with previous studies

No previous study gave detailed information on the incidence and clinical presentation of moderate to severe GO in relation to Graves' hyperthyroidism in a population, and the effect of a change in iodine intake level has not been studied.

Bartley et al. (25–28) retrospectively identified all patients in Olmsted County, MN, who had been diagnosed with GO between January 1976 and December 1990. Person × years of observation was approximately 1.5 million. A total of 120 cases (all grades of disease combined) were identified. All patients and nearly the entire population studied were of Caucasian ancestry. The overall incidence rate for women was 160 cases/million per year and for men 29 cases/million per year.

The present study included patients with incident moderate to severe GO in approximately 8.9 million persons × years of observation, and incidence rates were lower (women 26.7/million per year, men 5.4/million per year). Thus, our incidence values are approximately one sixth of those of Bartley et al. (25, 26) but with a similar sex ratio. A likely explanation for the difference in incidence rate is that the majority of GO patients in the study by Bartley et al. suffered from relatively mild GO. Thus, 42.5% of the patients in the Olmsted County study showed some sign of extraocular muscle affection (27), whereas this was 94.4% of patients in the present study. Bartley et al. (26) found higher incidence rates in some relatively narrow age groups (women 40–45 and 60–65 yr of age, men 45–50 and 65–70 yr of age). Our results did not confirm these findings, which may in part be caused by the difference in type of GO patients in the two studies (present study: all moderate to severe GO, Bartley study: mostly mild GO).

In a recent Swedish study on the incidence of hyperthyroidism, some clinical signs of GO were found in approximately 20% of patients with Graves' disease at the time of diagnosis of hyperthyroidism (3). If this fraction is applied to the study of Graves' hyperthyroidism in North Jutland, a GO (all grades of severity) incidence rate of 62/million per year would be expected.

The time relation between the development of Graves' hyperthyroidism and GO observed in the present study (Fig. 3) illustrates that the majority of patients with moderate to severe GO (63%) develops orbitopathy after developing Graves' hyperthyroidism. Thus, GO incidence figures obtained around the time of diagnosis of hyperthyroidism may be too low. Our time relation between Graves' hyperthyroidism and GO is quite similar to the results of the epidemiological study by Bartley et al. (28), in which 54% of patients developed GO after Graves' hyperthyroidism and 18% developed the diseases simultaneously, but different from the clinic based study of Wiersinga et al. (29). These authors found that only 37% of patients developed GO after Graves' hyperthyroidism, whereas 62% developed GO before or simultaneously with Graves' hyperthyroidism.

We found a relatively low incidence of GO in young (<40 yr of age) patients with Graves' disease. This is in accordance with the study of Perros et al. (30), and with the population based study of Bartley et al. (25, 26). On the other hand, Marcocci et al. (31) reported a similar risk of GO at all ages of patients with Graves' disease.

There was no difference in sex distribution between GO and Graves' hyperthyroidism in our study, which is at variance with several studies reporting a relative predominance of moderate to severe GO in men (30, 31). None of these other reports were regular population-based incidence studies.

Many classification systems for GO have been used (11–18). To evaluate disease activity, a much-used tool has been the Clinical Activity Score [CAS (15)] based on symptoms and signs of inflammation + history on disease development. During our study period from 1992 to 2009, we did not systematically register all components of CAS, such as, for example, caruncular edema. Most likely our scoring of inflammatory signs would correlate considerably to the inflammatory signs in CAS (15) and also to the more recent VISA Inflammatory Score (18). Concerning double vision, our system of recording did not include the details on history of intermittent diplopia used for grading according to Bahn and Gorman (14). Data given in the present study are based on the results of testing on the day of investigation.

Strengths and limitations

We prospectively studied all patients diagnosed with active moderate to severe GO in a well-defined population over a time period of 18 yr. The only other population-based study of GO is that of Bartley et al. (25–28). Many reports on GO have included the patients referred to a particular clinic with no information given on the underlying population or on the pattern of referral of patients diagnosed with GO in this population. Clinic-based studies of thyroid disease epidemiology may very well be biased from referral of only subgroups of patients (32), and referral patterns should be carefully evaluated. In the present study, we included only patients living in the primary catchment area of our Thyroid-Eye Clinic, which was the only center caring for patients with this disease in the area. In Denmark it would be very uncommon for patients developing abnormalities such as moderate to severe GO not to seek medical care, which is financed by the state. On the other hand, we cannot exclude that an old person with multiple diseases may have died with undiagnosed GO and without being included in the present study.

Definition of active moderate to severe GO is somewhat arbitrary, as discussed in the European Consensus report on GO (4). Both the American Thyroid Association and the American Association of Clinical Endocrinologists management guidelines on hyperthyroidism (33) and the European Consensus Report on GO (4) list a number of specific criteria that may guide in the classification of GO. They correspond in many details to criteria used in the present study, but because several details such as diplopia (as discussed above) have been registered differently, an exact comparison is not possible. However, clinical consequences are clear because patients with moderate to severe GO will be offered active immune modulating therapy (4), whereas patients with mild GO will not. We used the same criteria for diagnosing moderate to severe GO over the time period of our study. Most systems of classification include a systematic evaluation of corneal involvement. This complication to moderate to severe GO was not the primary cause for our classification of GO being moderate to severe in any patient.

Many studies of GO include patients with mild as well as moderate to severe disease, and it is often difficult to evaluate the clinical disease stage of patients included. Moreover, the delineation of very mild GO from no GO in a patient with Graves' hyperthyroidism may be difficult. Some studies have suggested that a large fraction of Graves' hyperthyroidism patients may suffer from subclinical GO (1).

We related our study of the epidemiology of moderate to severe GO to a recent study of the epidemiology of Graves' hyperthyroidism in nearly the same population. This allowed us to obtain reliable information on differences and similarities in sex and age distribution. No such comparative epidemiological study has been published previously.

Conclusion

Approximately 5% of patients with Graves' disease develop moderate to severe GO. The risk is similar in women and in men with Graves' disease but much higher in patients aged 40–60 yr than in young patients with Graves' disease. The overall incidence of moderate to severe GO was 16 cases/million per year, with no difference between periods before and after iodization of salt in Denmark.