Acute visual loss, attributable mostly to anterior ischemic optic neuropathy and less commonly to central retinal artery occlusion, and cerebrovascular accidents (CVAs) are among the leading causes of giant cell arteritis (GCA)–related morbidity (1–4). Other cranial ischemic complications such as tongue necrosis, scalp necrosis, and hearing loss are less common (5). Typically, visual loss and CVA occur suddenly, being the presenting manifestations of GCA in many of these cases. However, premonitory symptoms have been described in a substantial number of patients; these include transient ischemic attacks (TIAs), amaurosis fugax, blurry vision, transient diplopia, visual hallucinations, and jaw claudication (2, 6–10). Although most cranial ischemic complications occur at the time of presentation of GCA, they may develop soon after steroid therapy has begun, or later while the steroid dose is being tapered (2–4, 10–16).
Low-dose aspirin, due to its antiplatelet effect, has been shown to prevent ischemic events such as strokes and myocardial infarctions in patients with atherosclerotic cardiovascular disease (17, 18) and has been indicated for this purpose for 2 decades. The effect of low-dose aspirin on the rate of cranial ischemic complications in GCA had not been studied. In this retrospective study, we evaluated the effect of treatment with aspirin, prior to diagnosis of GCA and following commencement of steroid therapy, on the rate of cranial ischemic complications.
PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
Charts of all patients in whom GCA was diagnosed at the 4 main general hospitals in Jerusalem between 1980 and 2000 were reviewed. The diagnosis was biopsy-proven in 152 of these patients. Other patients (i.e., those with biopsy-negative GCA [n = 23]) were included if they met the 1990 American College of Rheumatology (ACR) criteria for the classification of GCA (19). Two patients in whom GCA was diagnosed prior to 1990 and who retrospectively did not meet these criteria were not included. Followup data were available from either outpatient clinic charts or the patients' family physicians; patients with <3 months of followup data were excluded. Altogether, admission and initial hospitalization data were available for 175 patients, and followup data were available for 166 of these patients.
Admission data included the patient's age, sex, symptoms at presentation, duration of symptoms, concomitant diseases and medications, significant findings of the physical examination (including ophthalmologic and neurologic examinations, when relevant), and results of laboratory tests (erythrocyte sedimentation rate and complete blood cell count). Followup data included the starting dose of prednisone, other medications used by the patient, and development of cranial ischemic complications. Patients presenting with cranial ischemic complications were diagnosed as having GCA if ACR criteria were fulfilled. Likewise, cranial ischemic complications developing within 14 days of the diagnosis of GCA were also considered to be GCA-related. Cranial ischemic complications that developed later, during tapering of the steroid dose or following discontinuation of steroids, were considered to be GCA-related only if they were associated with other GCA-related signs, symptoms, or laboratory evidence of an acute-phase reaction.
Associations between cranial ischemic complications, aspirin use, and all other explanatory variables were tested by Fisher's exact test. Variables associated with cranial ischemic complications with a level of significance of P < 0.09 were entered into a multivariate analysis. This higher level was determined in order to not omit borderline univariate associations that might prove significant when adjusting for other variables. Multivariate analysis was performed with logistic regression models. Analyses were performed using SPSS for Windows, version 11 (SPSS, Chicago, IL).
- Top of page
- PATIENTS AND METHODS
At the time of the GCA diagnosis, 36 patients (21%) had already been receiving low-dose aspirin therapy (100 mg/day). The indication for aspirin therapy in all cases was ischemic heart disease: 6 patients had had a myocardial infarction, and 30 had angina pectoris. Two of these patients also had had CVA at 2 and 3 years, respectively, prior to the GCA diagnosis. In Table 1, the characteristics of the aspirin-treated patients are compared with those of the 139 patients who were not treated with aspirin. There were no significant differences between the groups regarding the mean age of the patients, the male-to-female ratio, and several clinical and laboratory variables. Cerebrovascular risk factors (hypertension, hyperlipidemia, or diabetes) were more common in the aspirin-treated group (38.9% versus 20% in non–aspirin-treated patients) (P = 0.03).
Table 1. Characteristics of aspirin-treated and non–aspirin-treated patients with GCA at the time of diagnosis*
|Characteristic||Aspirin (n = 36)||No aspirin (n = 139)||P|
|Age, mean ± SD years||76.1 ± 7.5||73.5 ± 8.7||0.11|
|Headache||27 (75)||105 (76)||1.00|
|Systemic symptoms||25 (69)||73 (53)||0.09|
|Jaw claudication||6 (17)||25 (18)||1.00|
|Polymyalgia rheumatica||14 (39)||49 (35)||0.70|
|Symptom duration, weeks||6.5 ± 5.8||6.1 ± 6.6||0.22|
|Erythrocyte sedimentation rate, mm/hour||102 ± 27||103 ± 27||0.84|
|Platelet count, 103/μl||459 ± 154||448 ± 159||0.70|
|Hemoglobin, gm/dl||11.3 ± 1.5||11.2 ± 1.5||0.71|
|Positive temporal artery biopsy||34 (94)||118 (85)||0.17|
|Cerebrovascular risk factors||14 (38.9)||28 (20)||0.03|
|Cranial ischemic complications||3 (8)||40 (29)||0.01|
Despite the increased rate of risk factors, only 3 of the aspirin-treated patients (8%) presented with cranial ischemic complications, compared with 40 of the non–aspirin-treated patients (29%) (P = 0.01). After adjusting for covariates that were found to be significantly associated with cranial ischemic complications (male sex, jaw claudication, and lack of systemic symptoms and polymyalgia rheumatica), the odds ratio for cranial ischemic complications developing in patients who were already receiving aspirin at presentation relative to that in patients who had not been treated with aspirin was 0.22 (95% confidence interval [95% CI] 0.06–0.80, P = 0.02).
The nature of cranial ischemic complications at presentation is elaborated in Table 2. Transient ischemic attacks (TIA) or amaurosis fugax preceded cranial ischemic complications in 22% of these patients. Other GCA symptoms prior to the development of cranial ischemic complications were reported by most patients with CVA, with one-third of the patients having loss of vision. TIA or amaurosis fugax that did not develop into cranial ischemic complications was reported at presentation by 16 patients. Aspirin therapy did not affect the rate of these transient ischemic events: they occurred in 3 aspirin-treated patients (8%) and 13 patients not receiving aspirin (9%).
Table 2. CICs at the time of GCA diagnosis in 175 patients*
| ||Type of CIC|
|Visual loss (n = 32)||CVA (n = 13)|
|No. (%) of patients with other GCA symptoms prior to CICs||11 (34)||10 (77)|
|Mean duration (range) of other GCA symptoms prior to CICs, weeks||3 (0–28)||6 (0–24)|
|No. (%) of patients with premonitory amaurosis fugax or TIA||7 (22)||3 (23)|
|No. (%) of patients receiving aspirin treatment||2 (6)||1 (8)|
Following the diagnosis of GCA, all patients started therapy with prednisone. The starting dosage was 40–80 mg/day for most patients. Ten patients started with a dosage of ≤30 mg/day, and 3 patients started with a dosage of ≥100 mg/day. Treatment with aspirin was continued in all patients who had previously received aspirin therapy and was started in 41 additional patients within days of the diagnosis of GCA, because some of the treating physicians assumed it would be of benefit in this condition.
Nine patients were lost to followup (3 with prior aspirin therapy, 1 who started receiving aspirin following the diagnosis of GCA, and 5 who did not receive aspirin). Of the 166 remaining patients, 73 were receiving aspirin and prednisone, and 93 were receiving prednisone only. The mean followup time was similar in the 2 groups: 26.4 months in the aspirin-treated group, and 26.3 months in the patients treated with prednisone only. There were no significant differences between these 2 groups regarding the mean age of patients, the male-to-female ratio, several clinical and laboratory variables, and the mean starting dosage of prednisone (Table 3). Again, cerebrovascular risk factors were more common in the aspirin-treated group than in the non–aspirin-treated group (28.8% versus 12.9%) (P = 0.02).
Table 3. Characteristics of GCA patients with at least 3 months of followup*
|Characteristic||Aspirin (n = 73)||No aspirin (n = 93)||P|
|Age, mean ± SD years||75 ± 8.2||73.1 ± 8.7||0.16|
|Followup period, months||26.4 ± 22.8||26.3 ± 22.0||0.98|
|Headache||54 (74)||70 (75)||0.86|
|Systemic symptoms||39 (53)||53 (57)||0.75|
|Jaw claudication||17 (23)||12 (13)||0.10|
|Polymyalgia rheumatica||26 (36)||35 (38)||0.87|
|Erythrocyte sedimentation rate, mm/hour||102 ± 25||105 ± 27||0.39|
|Platelet count, 103/μl||451 ± 123||450 ± 181||0.94|
|Hemoglobin, gm/dl||11.4 ± 1.4||11.2 ± 1.6||0.37|
|Prednisone starting dosage, mg/day||53 ± 15||57 ± 16||0.18|
|Cerebrovascular risk factors||21 (28.8)||12 (12.9)||0.02|
|CICs during followup||2 (2.7)||12 (12.9)||0.02|
Cranial ischemic complications developed in 14 of these 166 patients during followup (Table 4). In 5 patients, these complications developed within 14 days of the initiation of steroids, and in the other 9 patients the development of these complications coincided with tapering or discontinuation of steroid therapy. Nine of the 14 patients (64%) had already experienced another cranial ischemic complication at presentation. Cranial ischemic complications developed in only 2 of the aspirin-treated patients (3%) compared with 12 of the non–aspirin-treated patients (13%) (P = 0.02). The odds ratio for cranial ischemic complications developing in aspirin-treated GCA patients following diagnosis and steroid therapy, relative to patients who did not receive aspirin, was 0.18 (95% CI 0.04–0.84, P = 0.03).
Table 4. CICs developing after GCA diagnosis and steroid treatment in 166 patients with at least 3 months of followup*
| ||Type of CIC|
|Visual loss (n = 8)||CVA (n = 6)|
|Mean duration of treatment prior to CIC (range)||8.3 months (2 days–30 months)||6.5 months (5 days–8 months)|
|No. of previous CICs at presentation||7 (all AION)||2 (1 AION, 1 CRAO)|
|No. of patients receiving aspirin||1||1|
- Top of page
- PATIENTS AND METHODS
The data presented here suggest that low-dose aspirin may decrease the incidence of cranial ischemic complications in patients with GCA. Patients receiving aspirin were 5 times less likely to experience cranial ischemic complications, either at presentation or during followup, compared with GCA patients who were not receiving aspirin therapy. In patients with GCA, cranial ischemic complications are among the presenting symptoms of the disease; thus, prevention of irreversible ischemia is difficult in such patients and may be achieved by increased awareness leading to early diagnosis of GCA. Nevertheless, there have been reports of ischemic complications developing in a substantial number of patients soon after GCA was diagnosed and treatment was started, or later during tapering of the dosage of steroids (2–4, 10–16). Therefore, patients with GCA may benefit from the addition of low-dose aspirin to the steroid treatment.
Data in this study were evaluated retrospectively and thus should be interpreted with caution. Because there was no predefined protocol, the results might have been biased in different directions. It is possible that some difference in a variable other than aspirin may explain the decreased incidence of cranial ischemic complications. Likely candidates for the increasing incidence of cranial ischemic complications are cerebrovascular risk factors such as hypertension, diabetes, hyperlipidemia, smoking, hyperhomocystinemia, antiphospholipid antibodies, or atrial fibrillation, while the use of anticoagulants or statins could potentially have a “protective” effect. Due to the retrospective nature of this study, we do not have complete information regarding some of these variables. Data on smoking, hyperhomocystinemia, antiphospholipid antibodies, and the use of statins were incomplete and thus were not included in the evaluation. Of the 3 patients with atrial fibrillation, 2 were receiving coumadin. None of these patients was receiving aspirin, and none experienced a cranial ischemic complication. No other patient was receiving coumadin or any other anticoagulant.
Data on hypertension, hyperlipidemia, and diabetes (the major cerebrovascular risk factors) were complete: the presence of these conditions was more common in the aspirin-treated group (Tables 1 and 3). Nevertheless, fewer cranial ischemic complications developed in aspirin-treated patients. Therefore, we believe these results are valid, because most patients were eventually evaluated and treated according to similar guidelines, and the difference in the odds ratio for cranial ischemic complications in aspirin-treated GCA patients, relative to GCA patients not treated with aspirin (after adjustment for the major potential confounders), was clearly significant.
Previous studies have suggested that the rate of cranial ischemic complications is inversely related to the intensity of the inflammatory response (10, 20, 21). Another study showed an increased rate of visual loss in patients with thrombocytosis (21). Patients with ischemia in other cranial tissues, such as that manifested by jaw claudication or scalp necrosis, are also considered to be at increased risk of developing irreversible cranial ischemic complications (10, 22). In our study the protective effect of aspirin remained significant after adjusting for such variables.
Cranial ischemic complications were reported at presentation in 25% of this group of GCA patients (visual loss in 18% and stroke in 7%). These values are comparable with the rates described in other studies (20, 21, 23). Regarding cranial ischemic complications developing after initiation of steroid therapy and during followup, the rates of visual loss and CVA in our patients were 5% and 4%, respectively.
Data on visual loss developing in GCA patients after initiation of steroid therapy are scarce, but the rate of visual loss in our patients was higher than that reported in other studies (10, 11). We suspect that one of the main reasons was a high rate of noncompliance, because 3 of the 8 patients in whom visual loss developed had decided to discontinue steroids prematurely due to side effects. It is possible that higher initial doses of steroids, especially in patients presenting with cranial ischemic complications, could have prevented some of the cases of visual loss developing soon after initiation of treatment. We are not aware of any epidemiologic data on GCA-related CVA developing after initiation of steroid therapy, although such patients have been described (10, 13, 24).
It is difficult at times to differentiate between arteritic and nonarteritic acute visual loss due to anterior ischemic optic neuropathy or central retinal artery occlusion, which may occur in conditions such as diabetes mellitus, hypertension, and hyperlipidemia. The presence of choroidal ischemia and the shape of the damage to the optic disc in cases of anterior ischemic optic neuropathy may help sometimes in differentiating between arteritic and nonarteritic etiology (2, 25, 26), but clinicians rely mostly on the presence of associated symptoms and signs and laboratory parameters of inflammation in order to make a diagnosis of arteritic anterior ischemic optic neuropathy or central retinal artery occlusion.
It is even more difficult to distinguish between atherosclerotic stroke and arteritic stroke. Although arteritic strokes tend to involve the vertebrobasilar system (24), this phenomenon is not a sufficient basis for diagnosis. Rather, the association of symptoms and signs suggestive of GCA, along with laboratory parameters of inflammation, is suggestive of a diagnosis of arteritic stroke. In this study, patients presenting initially with cranial ischemic complications were diagnosed as having GCA when the ACR criteria were fulfilled. Likewise, cranial ischemic complications that developed within 14 days of the diagnosis of GCA and the start of treatment were considered to be GCA-related. Cranial ischemic complications that developed during tapering of the steroid dosage or following discontinuation of steroids were considered to be GCA-related only if they were associated with other GCA-related symptoms or laboratory evidence of an inflammatory reaction.
Thus, episodes of CVA in 2 patients (both non–aspirin-treated) were considered nonarteritic, because neither the signs and symptoms of GCA nor laboratory evidence of inflammatory reaction were present at the time of the stroke. These 2 patients were not included in the final evaluation as having GCA-related strokes. It is possible that some of the other cases could have been misdiagnosed (being false-positive), but, according to the above-mentioned principles, we believe that such a possibility is negligible. Moreover, this finding of 2 observed cases is close to the expected number: a recent study showed the yearly age-standardized attack rate of stroke to be ∼3 per 1,000 (27). In our group of 166 patients, with an average followup period of 2.2 years, the expected number of strokes would be 1.
The lack of patients with nonarteritic anterior ischemic optic neuropathy in this population is also not surprising: it has been estimated that the annual incidence rate for nonarteritic anterior ischemic optic neuropathy is ∼10 per 100,000 (28). It can be calculated that for our group of 166 patients, more than 20 years of followup would be required to identify 1 case of anterior ischemic optic neuropathy.
The protective mechanism of aspirin in these patients is probably related to its antiplatelet effect. Although thrombosis is not reported as part of the histologic findings in temporal artery biopsies, it may be found in vertebral arteries with GCA at the site of inflammation and extending distally (29). It is possible that platelet thrombi in smaller, more distal inflamed arteries might cause the ischemic event. Recently, aspirin was reported to have action complementary to steroids in suppressing inflammation in GCA, predominantly by suppressing interferon-γ (30). The doses used in this mouse model of GCA were equivalent to 1–2 gm/day in humans. Data in our report suggest that even a lower dosage of 100 mg/day, probably acting as an antiplatelet rather than an antiinflammatory agent, is helpful in preventing GCA-related cranial ischemic complications.
Treatment with low-dose aspirin is not without risks: in a meta-analysis of 6 studies with 6,300 patients, 1 of 100 treated patients had nonfatal gastrointestinal (GI) bleeding (31). Steroids may also cause upper GI bleeding, so the combination of both medications may further increase the rate of upper GI bleeding. In our group of 73 aspirin-treated patients, only 1 (1.4%) had an episode of upper GI bleeding during the followup period. In comparison, 7 (7.5%) of the 93 patients treated with steroids developed this side effect. Low-dose aspirin may also affect the renal physiology, slightly decreasing creatinine clearance and uric acid excretion in elderly patients (32), but does not interfere with the blood pressure–lowering effect of antihypertensive agents (33).
Based on the data in this study, and the relatively favorable safety profile of low-dose aspirin, we suggest that addition of low-dose aspirin to steroid therapy should be considered in patients with GCA, at least for those presenting with GCA-related cranial ischemic complications, to minimize the chance for developing further cranial ischemic complications.