SEARCH

SEARCH BY CITATION

Keywords:

  • Ankylosing spondylitis;
  • Incidence;
  • Prevalence;
  • Survival

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Objective

To determine the incidence and prevalence of ankylosing spondylitis (AS) over a prolonged period in the 2 northernmost counties of Norway, where HLA–B27 has a high prevalence in the population.

Methods

We conducted a cohort study of all patients registered with a diagnosis of AS between 1960 and 1993 at the University Hospital of Northern Norway, which is the sole rheumatology department serving these counties. We registered demographics, year of disease onset (clinical disease), and year of diagnosis (radiograph confirmation) for all patients. The date of onset of clinical disease in patients with AS was used in the calculation of incidence rates. Annual incidence and point/period prevalence rates were expressed per 100,000 adults. Primary AS was defined as AS in the absence of psoriasis or inflammatory bowel disease (IBD).

Results

A total of 534 patients (75.1% male, mean age at clinical diagnosis 24.2 years, 93.0% HLA–B27 positive) had a confirmed diagnosis of AS (by the modified New York criteria). Median time from disease onset to radiologic confirmation was 8.0 years. Annual incidence of primary AS (n = 417) was 7.26, while estimated point prevalence rose from 0.036% in 1970 to 0.10% in 1980 and to 0.21% in 1990 with a period prevalence of 0.26%. AS was secondary to psoriasis or IBD in 117 patients (18.1%), with a diagnostic delay similar to that in primary AS. Annual incidence (14.1) and period prevalence in 1982–1993 (0.41%) were significantly higher in the town of Tromsø than in the surrounding rural region (5.21 and 0.22%, respectively). Mortality in patients with AS was low.

Conclusion

The incidence of AS was relatively stable in the northern part of Norway over a 34-year period. Incidence and prevalence are higher than reported in similar studies from Finland and Minnesota, possibly due to a higher population prevalence of HLA–B27.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Ankylosing spondylitis (AS) is a chronic inflammatory disease (1, 2) that primarily affects the sacroiliac joints (SI joints) and spine. The prevalence of AS shows considerable variation, with a 6% prevalence among the male population of the Haida Indians in Canada (3) and the virtual nonexistence of AS in several African populations (4–6). Although its cause remains elusive, an association with the HLA class I molecule B27 has been firmly established (7). The prevalence of AS correlates roughly with the population frequency of HLA–B27, supporting a role for HLA–B27-mediated antigen presentation in the pathogenesis of AS (8).

Studies on the incidence of AS are sparse and indicate an annual incidence of 6–7 per 100,000 persons in caucasian populations (9). Most studies on AS prevalence and incidence include patients with AS secondary to psoriasis, inflammatory bowel disease (IBD), and/or reactive arthritis. Because epidemiologic studies on primary AS in a region with a high population background of HLA–B27-positive individuals are rare, we performed a study based on hospital records to estimate the minimum incidence and prevalence of primary AS in the 2 northernmost counties of Norway.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Catchment area.

The Department of Rheumatology at the University Hospital of Northern Norway in Tromsø was founded in 1978 and remains the only department of rheumatology serving the counties Tromsø and Finnmark in Northern Norway. The department had a catchment area with an average population of 217,000 in the period 1960–1993, in which there was a slight decrease in the male:female ratio from 1.07 to 1.03, together with a decrease in the proportion of individuals under the age of 16 from 30.6% to 20.8% (these data have been extracted from the Web site of Statistics Norway at www.ssb.no). The region is rural and located ∼650 km north of the polar circle along the arctic sea, with Tromsø being the largest city with 60,000 citizens. The population consists mainly of caucasian individuals (Norsemen and Finnish immigrants constitute >96%) with an HLA–B27 prevalence of 15.9% (10, 11) and a Sami population of ∼10,000 with an HLA–B27 prevalence of ∼24% (12).

Patients.

Records for all patients registered in the hospital diagnostic database with a diagnosis of AS (International Classification of Diseases codes 720 and M45) since 1978 were reviewed. After evaluation of all patient records and radiographs of the SI joints, only patients fulfilling the modified New York criteria for AS (13) were included in the present study. In case of equivocal findings, a new radiologic examination of the SI joints was performed and patients without definite radiologic sacroiliitis were excluded. Data were recorded using a predefined data form retrieved from hospital records and supplemented during a reexamination of available patients in the outpatient clinic. The regional ethics committee approved the study protocol.

Methods.

We defined the time of onset of AS as the year of onset of symptoms. This time point was chosen rather than the time of diagnosis because it has been established that there is a significant diagnostic delay in patients with AS (14, 15), and also because we expected a significant increase in the number of patients diagnosed after 1978, when the department was founded. Using the time of diagnosis as the time of disease onset would therefore have introduced a considerable bias. In the estimation of the incidence rates, we excluded all patients with onset of symptoms before the age of 16. Patients with sacroiliitis secondary to psoriasis (clinical and/or histologic diagnosis) or IBD (endoscopy and/or histologic diagnosis) were classified as having secondary AS. All other patients were considered to have primary AS, and results were calculated separately for primary and secondary AS.

Statistics.

All data were stored in and analyzed with the use of a computerized database (Epi-Info version 6.01; Centers for Disease Control and Prevention, Atlanta, GA). Results are reported as median values unless indicated otherwise. Continuous data were analyzed by Mann-Whitney U test and contingency tables were analyzed by chi-square test or Fisher's exact test in case of low numbers. Incidence rates are given per 100,000 persons per year. Survival estimates were calculated by the Kaplan-Meier method. To estimate the risk of death for the AS cohort, we used the standardized mortality ratio, which compares the ratio of observed deaths with expected deaths in the Norwegian population. For controls, we used the mortality data from an earlier case-control study on mortality in rheumatoid arthritis from this same region and period (16). Resulting P values less than 0.05 were considered to indicate statistical significance.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

Demographics.

Of the total 687 patients (75.3% male) who fulfilled the inclusion criteria, 49 were residing outside the region at the time of symptom onset, and 104 (16%) reported disease onset before 1960 or after 1993. The remaining cohort consisted of 534 patients who were followed for at least 5 years after disease onset between 1960 and 1993. There was no significant difference in sex-age distribution, frequency of HLA–B27, or age at disease onset between the excluded and included patients (data not shown). Demographic data for the cohort are shown in Table 1. There was no significant difference in diagnostic delay between patients who were HLA–B27 positive and those who were negative (7.9 years versus 8.5 years; P = 0.9), but the mean age at symptom onset was significantly higher in the group of patients who were HLA–B27 negative (29.0 years versus 23.6 years; P = 0.01). Seventy-seven patients (12.1%) reported a juvenile (<16 years of age) onset of symptoms, and were therefore excluded from incidence analyses. A total of 117 patients (18.3%) were classified as having secondary AS (psoriasis in 79 patients, IBD in 44 patients, and both conditions in 6 patients). The prevalence of HLA–B27 was significantly higher in patients with primary AS (94.9%) than in patients with secondary AS (82.2%; P = 0.0005); however, there were no significant differences between patients with primary and those with secondary AS in terms of age at symptom onset, diagnostic delay, or sex distribution.

Table 1. Demographic data of the epidemiologic study cohort (n = 534) of patients with ankylosing spondylitis in Northern Norway*
CharacteristicsValue
  • *

    IBD = inflammatory bowel disease.

Male:female ratio3.1:1
Age at onset, mean ± SD years24.2 ± 8.5
HLA–B27 positive, %93.0
Diagnostic delay, mean (range) years8.0 (0–33)
Patients with psoriasis/IBD, no. (%)117 (18.1)

Incidence.

The estimated average annual incidence of primary AS in the adult population during the period of 1960–1993 was 7.26 (95% confidence interval [95% CI] 5.30–9.22). Annual AS incidence including secondary AS during that period was 8.71 (95% CI 6.38–11.04). In the period of 1960–1970, the incidence rate of primary AS was 5.38 (95% CI 3.54–7.22) per year, and the incidence rate was 6.52 (95% CI 4.56–8.48) per year when secondary cases were included. The corresponding incidence rates in the period 1971–1981 were 8.87 (95% CI 6.23–11.51) and 10.64 (95% CI 7.70–13.58), and in the period 1982–1993 the incidence rates were 7.17 (95% CI 5.94–9.82) and 8.59 (7.48–9.70) (Figure 1). The distribution of disease onset by single years in the period 1960–1993 (Figure 2) indicated higher incidence rates during 1971–1981. Also, the incidence of primary AS in the municipality of Tromsø in the period 1982–1993 was 14.1 (95% CI 11.15–17.05) per year, compared with 5.2 (95% CI 3.16–7.27) per year in the rest of the region (P < 0.0001) (Figure 3).

thumbnail image

Figure 1. Incidence of primary and secondary ankylosing spondylitis (AS) in Northern Norway in 1960–1993. Dark line indicates annual incidence of primary AS. Light line indicates annual incidence of primary and secondary AS.

Download figure to PowerPoint

thumbnail image

Figure 2. Distribution of disease onset by year in ankylosing spondylitis.

Download figure to PowerPoint

thumbnail image

Figure 3. Incidence of ankylosing spondylitis in the municipality of Tromsø compared with the remaining rural region.

Download figure to PowerPoint

Prevalence.

The period prevalence of primary and primary/secondary AS in the adult population for the period 1960–1993 was 0.26% and 0.31%, respectively. The point prevalence at January 1, 1970; January 1, 1980; and January 1, 1990 of primary AS was 0.036%, 0.10%, and 0.21%, respectively, and that of primary and secondary AS was 0.043%, 0.122%, and 0.26%, respectively (Figure 4). The point prevalence of primary and secondary AS in 1993 in the municipality of Tromsø was 0.41% versus 0.22% for the remaining region (P < 0.0001). The corresponding values in 1970, 1980, and 1990 were 0.10% versus 0.03%, 0.22% versus 0.10%, and 0.39% versus 0.22% (P < 0.000001).

thumbnail image

Figure 4. Changes in point prevalence of ankylosing spondylitis (AS) at defined time points during the period 1960–1993.

Download figure to PowerPoint

Survival.

Fifty-three patients in the entire cohort (n = 687) died during the observation period, with a patient fatality rate of 8.1% in 9,332 patient-years. Patient survival rates (Figure 5) at 10, 20, 30, and 35 years after the onset of symptoms were 99%, 98%, 92%, and 89%, respectively. Standardized mortality ratios (Table 2) showed an overall reduced mortality in patients with AS with no age- or sex-related differences.

thumbnail image

Figure 5. Patient survival estimates after onset of symptoms in 687 patients with ankylosing spondylitis in Northern Norway. Figures above the curve indicate number of patients still in study.

Download figure to PowerPoint

Table 2. Mortality in patients with ankylosing spondylitis (AS) and population controls according to age and sex*
 Age group, years
<4545–6465–80All
  • *

    SMR = standardized mortality ratio; NC = not calculated; 95% CI = 95% confidence interval.

  • Mortality rates per 1,000 patient years.

Women    
 Deaths (% of AS patients)0 (0)2 (2.8)2 (15.3)4
 Patient years5261,1673972,090
 AS group rate01.751.9
 Controls rate7.118.13919.1
 SMRNC0.090.130.10
 95% CINC0.03–0.440.12–0.320.02–0.45
Men    
 Deaths, n (% of AS patients)6 (3.2)19 (8.8)21 (36.8)46
 Patient years (No)1,9134,0631,2847,260
 AS group rate3.134.6716.35.7
 Controls rate8.138.850.328.5
 SMR0.380.120.320.2
 95% CI0.1–1.40.04–0.30.18–0.60.09–0.5

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES

In this region, the estimated annual incidence for primary AS in the period 1960–1993 was 7.3 (95% CI 5.3–9.2), and for combined primary and secondary AS it was 8.7 (95% CI 6.4–11.0) per 100,000 adults. There was no clear change in incidence estimates over the 34-year study period. The period prevalence of AS was 0.26% for primary AS and 0.31% for primary and secondary AS, whereas the point prevalence 6-doubled during the study period. Incidence and point prevalence were higher in the town of Tromsø than in the surrounding rural region.

Considering that AS is not an uncommon inflammatory disease, there are surprisingly few published data on the incidence of AS. A population-based study from Finland found an AS incidence rate of 6.9 per 100,000 person-years (2), and a survey based on hospital records in Rochester, Minnesota for the period 1934–1989 found the overall age- and sex-adjusted incidence rate of primary AS to be 6.3 per 100,000 person-years, and that of primary and secondary AS together to be 7.3 per 100,000 person-years (9). Except for the estimates in the town of Tromsø, our results are largely in agreement with these findings, despite some methodologic differences. The Finnish study excluded patients with psoriasis and reactive arthritis, but included patients with IBD under the assumption that patients with AS frequently have inflammatory gut lesions. However, we categorized patients with AS and psoriasis, Crohn's disease, or ulcerous colitis as having secondary AS, and included patients with AS secondary to reactive arthritis in the group with primary AS. The Finnish study (2) was based upon a national registry of patients entitled to receive specially reimbursed medication for AS, but no such registry exists in Norway. However, patients with AS who do not require or no longer require disease-specific pharmacologic treatment are not registered in this material, which could reflect a potential underestimate of AS incidence. However, the same group has conducted another survey on the incidence of rheumatic diseases, where the diagnosis of AS was confirmed by radiology, and found a similar figure for AS incidence (6.3 per 100,000 person-years) (17). Therefore, in spite of these differences in approach, the incidence of primary AS in Northern Norway based on hospitalized patients is comparable with the incidence of primary and secondary AS in Rochester, Minnesota and with the reported incidence of AS in Finland. In light of the higher prevalence of HLA–B27 and a previous estimate of AS prevalence in the region (11, 18), it is rather surprising that the incidence of AS calculated for the county of Tromsø was only 2 times higher than that reported in the US (9). The incidence of AS in northwestern Greece was estimated to be 1.5 per 100,000 person-years (19), and the estimated incidence of all types of spondylarthropathies in the Japanese population was 0.48 per 100,000 person years (20). These figures are considerably lower than the estimates from Finland, Minnesota, and our population, and could reflect the lower prevalence of HLA–B27.

The period prevalence of AS rose relatively steeply after 1978 in this study. Because this increase coincided with the establishment of the sole department of rheumatology in this region, the increased AS prevalence is most likely explained by the ensuing increased awareness of AS. This assumption is supported by the decrease in the median diagnostic delay from 11.0 years (range 0–33 years) in the period 1960–1970 to 5.0 years (range 0–16 years) in the period 1982–1993. Similarly, the data (Figure 2) also indicate a peak in the incidence rate of AS in the period around 1978, with lower rates in both the prior and ensuing periods. Rather than actual declining incidence rates, we believe this finding also represents the establishment of the rheumatology department. However, it is also possible that the knowledge about AS among the general practitioners in the region subsequently increased and that less severe cases were not referred to our department.

All of our estimates are based on data from the hospital registry and do not necessarily reflect all population cases; selection bias may have occurred. The extent of selection bias, however, is likely limited because there are only 3 additional local hospitals in this region, which refer most patients with rheumatic disease to our department, as do the majority of general practitioners in both counties. Although patients had free access to primary and secondary care, county authorities were responsible for all health care expenditures and required an evaluation by a specialist in the region before patients were allowed to receive medical care in other counties. Therefore, for financial reasons, most patients with active rheumatic diseases who were in need of specialist evaluation in this period have been registered at our department. In contrast, there was discrepancy between the incidence and prevalence rates in the town of Tromsø compared with the surrounding rural region, although patients from Tromsø and the rural region did not differ with regards to age at symptom onset, diagnostic delay, or sex distribution (data not shown). The fact that incidence rates were almost 3 times higher in the city of Tromsø could indicate that not all patients with AS in the more remote parts of the counties have been diagnosed and referred to our department. Increased awareness and knowledge of rheumatic diseases after the foundation of our department may have contributed to this, but in general the threshold for seeking medical attention has been found to be higher in rural areas, resulting in lower estimates of incidence and prevalence as seen with rheumatoid arthritis (21).

To validate our findings, we compared the number of patients being diagnosed at our clinic with the number of patients examined at the department of radiology with radiologic sacroiliitis. In 1990, 51 patients were diagnosed with arthritis after radiograph examination of the SI joints. In the same year, 26 patients received the diagnosis at our clinic. The department of radiology's statistical database does not allow us to examine all images individually, and among the 51 patients there could possibly be individuals already registered with the diagnosis of AS at an earlier stage. Also, the local AS society (personal communication) provided a rough estimate of 350 current patients with AS, which is in agreement with our prevalence estimates. Our data are based on the use of the modified New York criteria (13) because these were most relevant to the current study period. In contrast to more recent European Spondylarthropathy Study Group classification criteria (22), the New York criteria do not encompass patients with unspecified spondylarthropathies and therefore our results provide no more than a lowest common denominator for AS. In comparison with the population-based study by Gran et al (11) who estimated the prevalence of AS to be 1.1–1.4%, our estimates of incidence and prevalence lie at the lower end of this spectrum and should therefore be considered minimum incidence and prevalence rates in the region.

The survival rate in this AS cohort was 89% after 35 years, with the survival curve indicating a marked reduction in survival rates after 20–30 years of disease. Standardized mortality ratios for patients with AS (Table 2) indicated lower death rates in all age and sex subgroups of patients with AS compared with controls.

In conclusion, primary AS is a common disease, which currently affects at least 0.4% of the adult population in Northern Norway. Because these figures are hospital based, the factual prevalence might even be higher if rigorous population-based studies could be realized. A rather stable AS incidence rate combined with low mortality during the 20 years after disease onset resulted in a steadily increasing prevalence of AS.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES