Acta med. scand. Vol. 198 pp. 445-451, 1975
Sudden Death in Rheumatoid Arthritis with Atlanto-axial Dislocation Pawel Mikulowski, Frank A. Wollheim, Paul Rotmil and lngeborg Olsen From the Vcirnhem Hospital f o r Chronic Diseoses und the Divisi011 of Rheumntology, Deportment of Medicine, University of Lund, Mulmii General Hospital, Mulmo, Swleden
ABSTRACT. A post mortem material of 11 consecutive cases of severe atlanto-axial dislocation (a.a.d.) with cord compression is reported. The total number of deaths from rheumatoid arthritis (RA) during the period of 5 years was 104, and all were autopsied. Neurological symptoms correlated poorly to fatal a.a.d. Hemiplegia was found in three cases, one of which, however, was caused by thrombosis cerebri. Spastic signs were transiently recorded in two patients and dysphagia in a further two. Five patients had a history of recent vomiting. A.a.d. was the sole or main cause of death in 8 cases and contributory in 2. Sudden death occurred in 7 of the cases. Only 2 cases had obtained a correct diagnosis intra vitam. The CNS findings at autopsy consisted of cord compression (11/11 cases), cord malacia (2/11) and cerebral oedema (3/11). One case had polyarteritis and renal amyloidosis and one pulmonary carcinoma with metastatic spread. Signs of active inflammation in the axial joints were present in 4 cases. This study, based on systematic post mortem examinations, revealed an unexpectedly high and not previously reported incidence of fatal medulla compression in RA patients with a.a.d. (10%).
Atlanto-axial dislocation (a.a.d.) is a well known feature in rheumatoid arthritis (RA), occurring in not less than 25% of most hospital series (4, 13). The natural history in the latter retrospective study seems to indicate that on the whole the afflicted patients d o not have a higher mortality rate than RA cases without C 1-2 instability. Furthermore, spinal cord compression was only seen in 3 of 52 deaths. However, one weakness of this and other studies lies in the lack of consistent post mortem examination with a suitable technique enabling the visuali-
zation of the cervical spinal tract. We here report on 1 I cases of autopsy-proven cord compression, found during a limited period of 5 years in a general hospital for chronic diseases, indicating a greater ultimate importance of this condition than is generally appreciated. Clinical warning signs were sought retrospectively from the records. MATERIAL AND METHODS The cases emanated from a hospital for chronic disease serving a city population of approximately 250000 inhabitants. The number of beds and of RA patients cared for each year are shown in Table I. The autopsy rate was 90% and the autopsies were performed with a standardized technique, mostly by one and the same pathologist. The foramen occipitale region was examined in all cases regardless of clinical diagnosis. Cases of minor translocation of the odontoid process without obvious cord compression were not included in this study. I n 8 of the 11 cases blocks were removed containing the upper cervical spine and an adjacent part of the skull base. The specimens were fixed in formalin, sawed sagitally, decalcified and examined histologically. The compressed medullas and brains were examined by routine histological methods.
RESULTS Table I shows the number of all autopsies in the hospital during 1969-74 to be around 500 each year. The number of RA patients autopsied each year varied between 8 and 25, and was approximately 5 % of the total mortality in the hospital. All cases in this series had a severe form of disease and were in functional class 111 or I V on admission. The Waaler Rose test was positive in 9 cases and not recorded in the remaining two. The duration of the disease Acta med. scand. 198
Table I. Nutnber of lwspital beds, RA cases cared f . r rind nutopsies prrfortnrtl Autopsies Year
Beds
RA cases
I969 I970 1971 1972 1973 1974" Total
872 792 746 744 744 725
46 41 40 43 34 50
Total
RA cases
560 442 462 472 501 406 2 843
19 8 25 21 20 II 104
A.a.d. cases
I 2 I 4 1
2 II
10
Jan. 1 - Oct. 30.
a
20
LO 50 AGE YEARS
30
60
70
80
Fig. 1. Sex distribution, duration of disease and age at death in 1 1 cases of a.a.d.
was 12-45 years (mean 29.8) (Fig. 1). Ten of the patients were women. The age at death varied between 54 and 81 years (mean 68.3). Hospitalization at death had lasted from one month to 26 years. Six of the patients had been hospitalized for more than one year and 8 for more than 6 months. Neck pain at rest and on motion was recorded as pronounced in 7 cases, and in 4 of these (nos. 2 , 4 , 6 and 8) neck symptoms had been one of the presenting signs of RA. Neurologicrrl sigtis are summarized in Table 11. In 3 cases apparently no major neurological signs were recorded. Case 2 was treated under the diagnosis of thrombosis cerebri, but had atypical pains in the paralysed limbs. The vertigo in case 3 occurred in connection with indomethacin therapy. Case 7 had a number of episodes with clonic cramps and became disoriented a month prior to death. Case 9 had 2 months before death an episode with vomiting followed by generalized clonic cramps but no Babinski's sign. Case 10 suffered a humerus fracture I month prior to death, and was mentally
deranged from then on. She subsequently developed a flaccid hemiplegia with a transient Babinski's sign. Case l l had for many months swallowing difficulties and it was also noted that her voice had changed and become hoarse. Standrrrd X-roy examination of the neck had been performed in 4 of the cases, and although various degrees of erosion and sclerosis were found, in no instance was a.a.d. visible. In cases 2 and 9, X-ray examination including lateral views in forward flexion had been performed, and both showed unequivocal subluxation, the distance between dens and atlas being 15 and 8 mm, respectively. In case 10 X-ray examination was requested shortly before death but could not be performed due to the patient's poor condition. In the remaining 4 cases no X-rays of the neck were taken. Tertninal svtnptoms. Table I11 summarizes the dominating symptoms immediately preceding death. Seven of the patients died suddenly, mostly
Table 11. Neurologicul symptotns
Table 111. Terminal symptoms
Case no. I
2 3 4 5
6 7 8 9 10 I1
Duration before death Not reported Hemiplegia with pain 4 rno. Vertigo, suspicious extensor reflex 1 mo. Not reported Hemiplegia 1 mo. Not reported Disoriented, convulsions I mo. Dizziness 4 mo. Seizures 2 mo. Transient hemiplegia, d ysphagia 1 mo. Dysphagia, transient extensor reflex 14 mo.
1 2 3 4
5 6 7 8 9 10 11
Acra med. scand. 198
Case no. Breathing difficulties, sudden death Weakness, circulatory failure Sudden death while washing her face Respiratory arrest Weakness, fever, circulatory failure Abdominal pains, chills, vomiting Sudden death, vomiting Sudden death, vomiting Hypotension, slowly progressing coma Progressive coma Respiratory stridor, sudden death while asleep
-
Table IV. Macroscopic cind microscopic post
t w r t P r i t ,fiwlings
Gross anatomy Case no.
Medulla
Brain
Histology of axial joints
Other lesions
Oedema
Not examined
Renal concrements
3
Compression Compression with malacia Compression
Oedema Normal
4
Compression
Normal
Stomach ulcer Pyloric ulcer, polyarteritis, amyloidosis Ulcus scar in stomach
5
Compression
6 7
Compression Compression
Malacia of parietal lobe (unrelated) Congestion Normal
8
Compression
Congest ion
Not examined Active inflammation. rheumatic granulomas Non specific inflammation, fibrosis N o n specific inflammation, fibrosis Fibrosis Active inflammation, rheumatic granulomas, cancer metastases Fibrosis
9
Compression with malacia Compression Compression
Oederna Congestion Subtenorial oedema
Not examined Fibrosis, odontoid exostoses Fibrosis, odontoid exostoses
Psoriasis Stomach ulcer Rheumatic perimyocarditis
1
2
10 11
while resting in supine position. Five patients had experienced nausea and vomiting during the last week of life, but only one patient (no. 6) actually vomited in agony. In four patients death occurred after slowly progressing weakness. Post riiortein j7findi1ig.v. Table IV summarizes the post mortem findings. Common in all cases was displacement upwards and backwards of the odontoid process protruding into foramen magnum with impingement on the medulla (Fig. 2 ) . The protruded odontoid narrowed the spinal canal up to 4 mm sagitally (Fig. 3). The lower part of the medulla was subsequently caught and pressed against the posterior wall of foramen occipitale and
Fig. 2. Foramen magnum from above. O=apex of the odontoid, M=flattened medulla pressed against the posterior edge of the foramen. Case 8. Magnification x I .3.
Stomach ulcer, pancarditis Acute peritonitis Bronchial cancer with metastases Rheumatic carditis. acute stomach ulcer
adjacent vertebrae. The medulla was flattened and showed a compression furrow just below the olivae (Figs. 4 and 5). In case I 1 the displaced odontoid protruded 25 mm above the level of the foramen occipitale (Fig. 6) and compressed the medulla at the pontine border. In this case the membranes covering the odontoid were perforated. The compressed medullas seemed to be less vas-
F i g . 3. Sagittal section of the occipito-cervical block. O=odontoid pitted into foramen magnum, A =anterior arch of atlas, C=clivus. Atlanto-axial joint obliterated by fibrous tissue. Case 8. Scale 75%. A r m med. srund. 198
448 P. Mikulorr3ski r t (11.
-.
Fig. 4 . Asymmetric, depressed medulla ( C ) Cafe 8.
Magnification ~ 0 . 9 .
cular; the histological lesions were, however, only indistinct with different stages of nuclear damage in ganglionic cells, mostly on the ventral side of the medulla. Except for cases 2 and 9, showing recent malacia, no haemorrhages or inflammation were observed at the compressed areas. The distance between the odontoid process and the anterior atlas arch in formalin-fixed, sagitallysawn specimens was increased up to 10 mm (Fig. 7). The space was filled with connective tissue, extending irregularly around the odontoid, atlas and adjacent occipital bone. The displacement of the odontoid was also due to fibrous ingrowth of the apophyseal joints between CI and C2. Histologically, destruction of articular surfaces of the odontoid was prominent. The damage was due to cartilage disappearance and ingrowth of fibrous
Fig. 5 . Histological section of the compressed medulla. C=depression on the ventral surface of the medulla. Case 8. Magnification x4.
pannus. The ingrowing fibrous tissue distended and obliterated articular spaces, leaving only some clefts lined by mesothelial cells (Fig. 7). The bones were osteoporotic with a mostly inconspicuous cortical layer, producing trabecular outgrowths. Fibrous tissue replacing the bone marrow merged
Fig. 7 . Histological section of sagitally transsected speciFig. 6 . Odontoid process (0) herniated 25 mm above the level of the foramen magnum; C=clivus. Case 1 I . Scale
( A ) , odontoid (0)and clivus ( C ) ;erosions on the cortical
63%.
layer of the odontord. Case 10. Magnification x 1.5.
Acta med. scand. 198
men. Dense fibrous tissue between anterior arch of atlas
Rheurnatoid arthritis with atlanto-axial dislocation
Fig. 8. Rheumatic granuloma in the obliterated atlanto-
axial joint. Note irregularly eroded contours of the adjacent cortical bone of the odontoid. Case 7. Magnifica-
449
Fig. 9. Rheumatic granuloma in Fig. 8 magnified x252.
tion x97.5.
with tissues outside the bones. The fibrous tissue in cases 4 and 10 was focally infiltrated by plasma cells and lymphocytes. In cases 3 and 7 there were, moreover. specific rheumatic granulomas with fibrinoid necroses surrounded by palissaded histiocytic cells (Figs. 8 and 9). Ciri4ses of death. Analysis of autopsy findings a s related to clinical symptoms unequivocally implicate a.a.d. with cord compression as the only cause of death in 5 cases (Table V ) . In the remaining cases other causes of death must also be considered. Thus in case 5 a thrombotic occlusion of the medial cerebral artery with a corresponding malacia of the parietal lobe was the obvious cause of death. Case 3 had a moderate uraemia with a stable serum creatinine of 4 mg/100 ml, and renal amyloidosis was found to be the underlying cause. The unexpected sudden death of the patient was, however,
judged to be due to spinal cord compression. In case 6 a gangrenous lesion in the distal part of ileum had perforated. In case 7, neurologic symptoms together with sudden death led to the clinical suspicion of intracranial metastasizing carcinoma. However, the post mortem examination showed no CNS tumor growth. and the immediate cause of death was considered to be the a.a.d. In case 10 the necropsy showed ;in acute pyelonephritis and the infection was listed as the main cause of death and the cord compression as contributory.
DISCUSSION
The present series of fatal atlanto-axial dislocation stresses the diagnostic difficulties encountered when dealing with severely disabled patients with rheumatoid arthritis. In only two of I I cases was the dislocation diagnosed before death and in one more case it was suspected shortly prior to death. On retrospective analysis of neurological symptoms and signs, only non-specific patterns could be found Table V . Causes of deiith and three cases apparently had no neurological findings at all. Spastic symptoms were only noted in Case Main cause Accessory cause no. ofdeath of death three o r four of the patients. These results are in accord with published clinical reports of systemati1 a.a.d. cally examined hospital patients with RA (4). Thus 2 a.a.d. 3 a.a.d. Renal amyloidosis neurological signs are not helpful for diagnosing 4 a.a.d. patients with a.a.d. o r pointing out those at risk of 5 Cerebral artery developing fatal cord compression. thrombosis a.a.d.? 6 Peritonitis a.a.d. Whereas the incidence of a.a.d is accurately 7 a.a.d. Pulmonary adenocarcinoma known from earlier studies, only indirect and unre8 a.a.d. Rheumatic carditis liable data are reported regarding medullary com9 a.a.d. 10 Acute pyelonephritis a.a.d. pression caused by a.a.d. These patients die often II a.a.d. unexpectedly with a clinical picture compatible
29-752986
Acta med. scand. 198
450
P. Mikulowski et al.
with, for instance, such common conditions as cerebrovascular accidents, myocardial infarction or pulmonary embolism. Autopsy rates may be low and even when autopsy is performed, one may miss the condition if skull opening and removal of the brain is performed without close observation of the foramen magnum and upper cervical region. Consequently autopsy reports on spontaneous fatal cord compression in RA are very few and consist of 6 cases ( I , S , 7 , 8 , 14, 16). In one of these cases both vertebral arteries were thrombosed. In addition, Hauge (6) described two deaths occurring after attempted surgical decompression. Discussion of a.a.d. not related to R A is beyond the scope of this paper; for references see Davis and Markely ( 5 ) and Wadia ( IS). Many R A patients die at home or in asylum and post mortem examinations are seldom performed. This situation is illustrated by Smith et al. (13) describing a group of 130 patients with clinically observed rheum. cervical luxation. In only one of 46 patients dying during the follow-up period was there a clinical suggestion of fatal cord compression: however, in no case was a post mortem examination performed. These authors conclude that the luxation per se in the absence of cord or brainstem involvement does not shorten survival. However, one must question the validity of their statement on the ground that post mortem studies on control and luxation cases are incomplete. Fatal cord compression deaths could be hidden in the both groups. The study of Mathews (10) also suffers from the lack of autopsy data on the deceased patients. Werne ( 17) established the dependence of atlantoaxial stability upon intact transverse and alar ligaments, and Ball and Sharp ( I ) also stressed the importance of osteoporotic changes and erosions caused by granulation tissue in the weakening of these ligaments. Our specimens were sawn sagitally in order to visualize odontoid protrusion. This method precluded gross examination of the ligaments, which in the histological sections were scarcely discernible in the fibrous tissues surrounding the odontoid. One histological finding which may be of significance is that of rheumatoid granulomas in two of our cases as well as in two of those published by others (5, 8). indicating a rather malignant type of RA. Fatal impingement on the medulla seems to be a final event in a slowly progressing displacement of Acta med. scand. 198
the odontoid process. The death can be precipitated by sudden head movements such as induced by vomiting. However, an eliciting factor could not always be identified. As only minimal lesions could be observed in the Compressed medullas, the death is apparently due to sudden ischemia of the reticular substance, containing respiratory and circulatory centra. This supposition is corroborated by cases 2 and 9 developing malacia of the medulla, with death coming less suddenly. We conclude that a.a.d. occurs in approximately 10% of RA patients dying in a Swedish nursing home population. Assuming a total incidence of roentgenologically detectable a.a.d. of 30% among these patients, the mortality rate would be around 25 5%. In view of the encouraging results with surgical fusion reported recently by Brattstrom and Granholm (2, 3), one must consider this therapeutic possibility, even in cases without cord or brain stem symptoms, especially since the majority of the patients suffer from pains i n the neck region. Cervical collars are often poorly tolerated and ineffective as stabilizers. The role played by corticosteroid treatment in the development of a.a.d. is controversial. Smith et al. (13) found an adverse effect on the progression. Among our patients 3 had received no steroids and 2 had steroids only for brief periods and in low doses. These observations do not support a major importance of steroid treatment for the development of fatal a.a.d. A s pointed o u t by Werne (It%), Robinson (12) and Ornilla et al. ( 1 I), a.a.d. has to be considered as a possible risk in connection with anesthesia for surgery in RA. Physiotherapeutic manipulations may be dangerous, too.
REFERENCES I . Ball, J. & Sharp, J.: Rheumatoid arthritis of the cervical spine. In: Modern trends in rheumatology, 2nd ed. p. 117. AGS Hill, Butterworth. London 1971. 2 . Brattstrom, H . & Granholm. L.: Chirurgie der Halswirbelsaule bei Patienten mit rheumatoider Arthitis. Orthopade 2: 118, 1973. 3. - Operative treatment of atlanto-axial luxation in rheumatoid arthritis. In manuscript 1975. 4. Conlon, P. W . , Isdale, I . C. & Rose, 8 . S.:
Rheumatoid arthritis of the cervical Spine Ann. rheum. Dis. 25: 120, 1966. 5 . Davis, Jr, F. W. & Markley, H. E.: Rheumatoid arthritis with death from medullary compression. Ann. Med. 35: 451, 1951.
6. Hauge, T.: So called spontaneous cervical dislocations. Acta chir. scand., Suppl. 232: 5 , 1958. 7. Martel, E. & Abel, M. R.: Fatal atlanto-axial luxation in rheumatoid arthritis. Arthr. and Rheum. 6: 224. 1963. 8. Martel, W . & Page, J . W.: Cervical vertebral erosions and subluxations in rheumatoid arthritis and ankylosing spondylitis. Arthr. and Rheum. 3: 546. 1960. 9. Mathews, J . : Atlanto-axial subluxation in rheumatoid arthritis Ann. rheum. Dis. 28: 260. 1969. 10. Mathews, J . A,: Atlanto-axial subluxation in rheumatoid arthritis. A 5-year follow-up study. Ann. rheum. Dis. 33: 526, 1974. 1 1 . Ornilla. E., Ansell, B. M. & Swannell. A. J.: Cervical spine involvement in patients with chronic arthritis undergoing orthopedic surgery. Ann. rheum. Dis. 3 I : 364, 1972.
12. Robinson. H. S.: Rheumatoid arthritis-atlanto-axial subluxation and its clinical presentation. Canad. med. Ass. J. 94: 470, 1966. 13. Smith, P. H . , Benn. R. T. & Sharp, J.: Natural history of rheumatoid cervical luxations. Ann. rheum. Dis. 31: 431, 1972. 14. Storey. G.: Changes in the cervical spine in rheumatoid arthritis with compressions of the cord. Ann. phys. Med. 4: 216, 1958. 15. Wadia, N . H.: Myelopathy complicating congenital atlanto-axial dislocation. Brain 9 0 449, 1967. 16. Webb, F. W. S . . Hickman. J. A. & Brew, D. St. J . : Death from vertebral artery thrombosis in rheumatoid arthritis. Brit. med. J. 2: 537, 1968. 17. Werne, S . : Studies in spontaneous atlas dislocation. Acta orthop. scand.. Suppl. 23: 5, 1957. 18. - Narkosrisken vid atlasdislokation. Lakartidningen 64: 4475. 1967.
Acta med. scand. 198
Sudden Death in Rheumatoid Arthritis with Atlanto-axial Dislocation Pawel Mikulowski, Frank A. Wollheim, Paul Rotmil and lngeborg Olsen From the Vcirnhem Hospital f o r Chronic Diseoses und the Divisi011 of Rheumntology, Deportment of Medicine, University of Lund, Mulmii General Hospital, Mulmo, Swleden
ABSTRACT. A post mortem material of 11 consecutive cases of severe atlanto-axial dislocation (a.a.d.) with cord compression is reported. The total number of deaths from rheumatoid arthritis (RA) during the period of 5 years was 104, and all were autopsied. Neurological symptoms correlated poorly to fatal a.a.d. Hemiplegia was found in three cases, one of which, however, was caused by thrombosis cerebri. Spastic signs were transiently recorded in two patients and dysphagia in a further two. Five patients had a history of recent vomiting. A.a.d. was the sole or main cause of death in 8 cases and contributory in 2. Sudden death occurred in 7 of the cases. Only 2 cases had obtained a correct diagnosis intra vitam. The CNS findings at autopsy consisted of cord compression (11/11 cases), cord malacia (2/11) and cerebral oedema (3/11). One case had polyarteritis and renal amyloidosis and one pulmonary carcinoma with metastatic spread. Signs of active inflammation in the axial joints were present in 4 cases. This study, based on systematic post mortem examinations, revealed an unexpectedly high and not previously reported incidence of fatal medulla compression in RA patients with a.a.d. (10%).
Atlanto-axial dislocation (a.a.d.) is a well known feature in rheumatoid arthritis (RA), occurring in not less than 25% of most hospital series (4, 13). The natural history in the latter retrospective study seems to indicate that on the whole the afflicted patients d o not have a higher mortality rate than RA cases without C 1-2 instability. Furthermore, spinal cord compression was only seen in 3 of 52 deaths. However, one weakness of this and other studies lies in the lack of consistent post mortem examination with a suitable technique enabling the visuali-
zation of the cervical spinal tract. We here report on 1 I cases of autopsy-proven cord compression, found during a limited period of 5 years in a general hospital for chronic diseases, indicating a greater ultimate importance of this condition than is generally appreciated. Clinical warning signs were sought retrospectively from the records. MATERIAL AND METHODS The cases emanated from a hospital for chronic disease serving a city population of approximately 250000 inhabitants. The number of beds and of RA patients cared for each year are shown in Table I. The autopsy rate was 90% and the autopsies were performed with a standardized technique, mostly by one and the same pathologist. The foramen occipitale region was examined in all cases regardless of clinical diagnosis. Cases of minor translocation of the odontoid process without obvious cord compression were not included in this study. I n 8 of the 11 cases blocks were removed containing the upper cervical spine and an adjacent part of the skull base. The specimens were fixed in formalin, sawed sagitally, decalcified and examined histologically. The compressed medullas and brains were examined by routine histological methods.
RESULTS Table I shows the number of all autopsies in the hospital during 1969-74 to be around 500 each year. The number of RA patients autopsied each year varied between 8 and 25, and was approximately 5 % of the total mortality in the hospital. All cases in this series had a severe form of disease and were in functional class 111 or I V on admission. The Waaler Rose test was positive in 9 cases and not recorded in the remaining two. The duration of the disease Acta med. scand. 198
Table I. Nutnber of lwspital beds, RA cases cared f . r rind nutopsies prrfortnrtl Autopsies Year
Beds
RA cases
I969 I970 1971 1972 1973 1974" Total
872 792 746 744 744 725
46 41 40 43 34 50
Total
RA cases
560 442 462 472 501 406 2 843
19 8 25 21 20 II 104
A.a.d. cases
I 2 I 4 1
2 II
10
Jan. 1 - Oct. 30.
a
20
LO 50 AGE YEARS
30
60
70
80
Fig. 1. Sex distribution, duration of disease and age at death in 1 1 cases of a.a.d.
was 12-45 years (mean 29.8) (Fig. 1). Ten of the patients were women. The age at death varied between 54 and 81 years (mean 68.3). Hospitalization at death had lasted from one month to 26 years. Six of the patients had been hospitalized for more than one year and 8 for more than 6 months. Neck pain at rest and on motion was recorded as pronounced in 7 cases, and in 4 of these (nos. 2 , 4 , 6 and 8) neck symptoms had been one of the presenting signs of RA. Neurologicrrl sigtis are summarized in Table 11. In 3 cases apparently no major neurological signs were recorded. Case 2 was treated under the diagnosis of thrombosis cerebri, but had atypical pains in the paralysed limbs. The vertigo in case 3 occurred in connection with indomethacin therapy. Case 7 had a number of episodes with clonic cramps and became disoriented a month prior to death. Case 9 had 2 months before death an episode with vomiting followed by generalized clonic cramps but no Babinski's sign. Case 10 suffered a humerus fracture I month prior to death, and was mentally
deranged from then on. She subsequently developed a flaccid hemiplegia with a transient Babinski's sign. Case l l had for many months swallowing difficulties and it was also noted that her voice had changed and become hoarse. Standrrrd X-roy examination of the neck had been performed in 4 of the cases, and although various degrees of erosion and sclerosis were found, in no instance was a.a.d. visible. In cases 2 and 9, X-ray examination including lateral views in forward flexion had been performed, and both showed unequivocal subluxation, the distance between dens and atlas being 15 and 8 mm, respectively. In case 10 X-ray examination was requested shortly before death but could not be performed due to the patient's poor condition. In the remaining 4 cases no X-rays of the neck were taken. Tertninal svtnptoms. Table I11 summarizes the dominating symptoms immediately preceding death. Seven of the patients died suddenly, mostly
Table 11. Neurologicul symptotns
Table 111. Terminal symptoms
Case no. I
2 3 4 5
6 7 8 9 10 I1
Duration before death Not reported Hemiplegia with pain 4 rno. Vertigo, suspicious extensor reflex 1 mo. Not reported Hemiplegia 1 mo. Not reported Disoriented, convulsions I mo. Dizziness 4 mo. Seizures 2 mo. Transient hemiplegia, d ysphagia 1 mo. Dysphagia, transient extensor reflex 14 mo.
1 2 3 4
5 6 7 8 9 10 11
Acra med. scand. 198
Case no. Breathing difficulties, sudden death Weakness, circulatory failure Sudden death while washing her face Respiratory arrest Weakness, fever, circulatory failure Abdominal pains, chills, vomiting Sudden death, vomiting Sudden death, vomiting Hypotension, slowly progressing coma Progressive coma Respiratory stridor, sudden death while asleep
-
Table IV. Macroscopic cind microscopic post
t w r t P r i t ,fiwlings
Gross anatomy Case no.
Medulla
Brain
Histology of axial joints
Other lesions
Oedema
Not examined
Renal concrements
3
Compression Compression with malacia Compression
Oedema Normal
4
Compression
Normal
Stomach ulcer Pyloric ulcer, polyarteritis, amyloidosis Ulcus scar in stomach
5
Compression
6 7
Compression Compression
Malacia of parietal lobe (unrelated) Congestion Normal
8
Compression
Congest ion
Not examined Active inflammation. rheumatic granulomas Non specific inflammation, fibrosis N o n specific inflammation, fibrosis Fibrosis Active inflammation, rheumatic granulomas, cancer metastases Fibrosis
9
Compression with malacia Compression Compression
Oederna Congestion Subtenorial oedema
Not examined Fibrosis, odontoid exostoses Fibrosis, odontoid exostoses
Psoriasis Stomach ulcer Rheumatic perimyocarditis
1
2
10 11
while resting in supine position. Five patients had experienced nausea and vomiting during the last week of life, but only one patient (no. 6) actually vomited in agony. In four patients death occurred after slowly progressing weakness. Post riiortein j7findi1ig.v. Table IV summarizes the post mortem findings. Common in all cases was displacement upwards and backwards of the odontoid process protruding into foramen magnum with impingement on the medulla (Fig. 2 ) . The protruded odontoid narrowed the spinal canal up to 4 mm sagitally (Fig. 3). The lower part of the medulla was subsequently caught and pressed against the posterior wall of foramen occipitale and
Fig. 2. Foramen magnum from above. O=apex of the odontoid, M=flattened medulla pressed against the posterior edge of the foramen. Case 8. Magnification x I .3.
Stomach ulcer, pancarditis Acute peritonitis Bronchial cancer with metastases Rheumatic carditis. acute stomach ulcer
adjacent vertebrae. The medulla was flattened and showed a compression furrow just below the olivae (Figs. 4 and 5). In case I 1 the displaced odontoid protruded 25 mm above the level of the foramen occipitale (Fig. 6) and compressed the medulla at the pontine border. In this case the membranes covering the odontoid were perforated. The compressed medullas seemed to be less vas-
F i g . 3. Sagittal section of the occipito-cervical block. O=odontoid pitted into foramen magnum, A =anterior arch of atlas, C=clivus. Atlanto-axial joint obliterated by fibrous tissue. Case 8. Scale 75%. A r m med. srund. 198
448 P. Mikulorr3ski r t (11.
-.
Fig. 4 . Asymmetric, depressed medulla ( C ) Cafe 8.
Magnification ~ 0 . 9 .
cular; the histological lesions were, however, only indistinct with different stages of nuclear damage in ganglionic cells, mostly on the ventral side of the medulla. Except for cases 2 and 9, showing recent malacia, no haemorrhages or inflammation were observed at the compressed areas. The distance between the odontoid process and the anterior atlas arch in formalin-fixed, sagitallysawn specimens was increased up to 10 mm (Fig. 7). The space was filled with connective tissue, extending irregularly around the odontoid, atlas and adjacent occipital bone. The displacement of the odontoid was also due to fibrous ingrowth of the apophyseal joints between CI and C2. Histologically, destruction of articular surfaces of the odontoid was prominent. The damage was due to cartilage disappearance and ingrowth of fibrous
Fig. 5 . Histological section of the compressed medulla. C=depression on the ventral surface of the medulla. Case 8. Magnification x4.
pannus. The ingrowing fibrous tissue distended and obliterated articular spaces, leaving only some clefts lined by mesothelial cells (Fig. 7). The bones were osteoporotic with a mostly inconspicuous cortical layer, producing trabecular outgrowths. Fibrous tissue replacing the bone marrow merged
Fig. 7 . Histological section of sagitally transsected speciFig. 6 . Odontoid process (0) herniated 25 mm above the level of the foramen magnum; C=clivus. Case 1 I . Scale
( A ) , odontoid (0)and clivus ( C ) ;erosions on the cortical
63%.
layer of the odontord. Case 10. Magnification x 1.5.
Acta med. scand. 198
men. Dense fibrous tissue between anterior arch of atlas
Rheurnatoid arthritis with atlanto-axial dislocation
Fig. 8. Rheumatic granuloma in the obliterated atlanto-
axial joint. Note irregularly eroded contours of the adjacent cortical bone of the odontoid. Case 7. Magnifica-
449
Fig. 9. Rheumatic granuloma in Fig. 8 magnified x252.
tion x97.5.
with tissues outside the bones. The fibrous tissue in cases 4 and 10 was focally infiltrated by plasma cells and lymphocytes. In cases 3 and 7 there were, moreover. specific rheumatic granulomas with fibrinoid necroses surrounded by palissaded histiocytic cells (Figs. 8 and 9). Ciri4ses of death. Analysis of autopsy findings a s related to clinical symptoms unequivocally implicate a.a.d. with cord compression as the only cause of death in 5 cases (Table V ) . In the remaining cases other causes of death must also be considered. Thus in case 5 a thrombotic occlusion of the medial cerebral artery with a corresponding malacia of the parietal lobe was the obvious cause of death. Case 3 had a moderate uraemia with a stable serum creatinine of 4 mg/100 ml, and renal amyloidosis was found to be the underlying cause. The unexpected sudden death of the patient was, however,
judged to be due to spinal cord compression. In case 6 a gangrenous lesion in the distal part of ileum had perforated. In case 7, neurologic symptoms together with sudden death led to the clinical suspicion of intracranial metastasizing carcinoma. However, the post mortem examination showed no CNS tumor growth. and the immediate cause of death was considered to be the a.a.d. In case 10 the necropsy showed ;in acute pyelonephritis and the infection was listed as the main cause of death and the cord compression as contributory.
DISCUSSION
The present series of fatal atlanto-axial dislocation stresses the diagnostic difficulties encountered when dealing with severely disabled patients with rheumatoid arthritis. In only two of I I cases was the dislocation diagnosed before death and in one more case it was suspected shortly prior to death. On retrospective analysis of neurological symptoms and signs, only non-specific patterns could be found Table V . Causes of deiith and three cases apparently had no neurological findings at all. Spastic symptoms were only noted in Case Main cause Accessory cause no. ofdeath of death three o r four of the patients. These results are in accord with published clinical reports of systemati1 a.a.d. cally examined hospital patients with RA (4). Thus 2 a.a.d. 3 a.a.d. Renal amyloidosis neurological signs are not helpful for diagnosing 4 a.a.d. patients with a.a.d. o r pointing out those at risk of 5 Cerebral artery developing fatal cord compression. thrombosis a.a.d.? 6 Peritonitis a.a.d. Whereas the incidence of a.a.d is accurately 7 a.a.d. Pulmonary adenocarcinoma known from earlier studies, only indirect and unre8 a.a.d. Rheumatic carditis liable data are reported regarding medullary com9 a.a.d. 10 Acute pyelonephritis a.a.d. pression caused by a.a.d. These patients die often II a.a.d. unexpectedly with a clinical picture compatible
29-752986
Acta med. scand. 198
450
P. Mikulowski et al.
with, for instance, such common conditions as cerebrovascular accidents, myocardial infarction or pulmonary embolism. Autopsy rates may be low and even when autopsy is performed, one may miss the condition if skull opening and removal of the brain is performed without close observation of the foramen magnum and upper cervical region. Consequently autopsy reports on spontaneous fatal cord compression in RA are very few and consist of 6 cases ( I , S , 7 , 8 , 14, 16). In one of these cases both vertebral arteries were thrombosed. In addition, Hauge (6) described two deaths occurring after attempted surgical decompression. Discussion of a.a.d. not related to R A is beyond the scope of this paper; for references see Davis and Markely ( 5 ) and Wadia ( IS). Many R A patients die at home or in asylum and post mortem examinations are seldom performed. This situation is illustrated by Smith et al. (13) describing a group of 130 patients with clinically observed rheum. cervical luxation. In only one of 46 patients dying during the follow-up period was there a clinical suggestion of fatal cord compression: however, in no case was a post mortem examination performed. These authors conclude that the luxation per se in the absence of cord or brainstem involvement does not shorten survival. However, one must question the validity of their statement on the ground that post mortem studies on control and luxation cases are incomplete. Fatal cord compression deaths could be hidden in the both groups. The study of Mathews (10) also suffers from the lack of autopsy data on the deceased patients. Werne ( 17) established the dependence of atlantoaxial stability upon intact transverse and alar ligaments, and Ball and Sharp ( I ) also stressed the importance of osteoporotic changes and erosions caused by granulation tissue in the weakening of these ligaments. Our specimens were sawn sagitally in order to visualize odontoid protrusion. This method precluded gross examination of the ligaments, which in the histological sections were scarcely discernible in the fibrous tissues surrounding the odontoid. One histological finding which may be of significance is that of rheumatoid granulomas in two of our cases as well as in two of those published by others (5, 8). indicating a rather malignant type of RA. Fatal impingement on the medulla seems to be a final event in a slowly progressing displacement of Acta med. scand. 198
the odontoid process. The death can be precipitated by sudden head movements such as induced by vomiting. However, an eliciting factor could not always be identified. As only minimal lesions could be observed in the Compressed medullas, the death is apparently due to sudden ischemia of the reticular substance, containing respiratory and circulatory centra. This supposition is corroborated by cases 2 and 9 developing malacia of the medulla, with death coming less suddenly. We conclude that a.a.d. occurs in approximately 10% of RA patients dying in a Swedish nursing home population. Assuming a total incidence of roentgenologically detectable a.a.d. of 30% among these patients, the mortality rate would be around 25 5%. In view of the encouraging results with surgical fusion reported recently by Brattstrom and Granholm (2, 3), one must consider this therapeutic possibility, even in cases without cord or brain stem symptoms, especially since the majority of the patients suffer from pains i n the neck region. Cervical collars are often poorly tolerated and ineffective as stabilizers. The role played by corticosteroid treatment in the development of a.a.d. is controversial. Smith et al. (13) found an adverse effect on the progression. Among our patients 3 had received no steroids and 2 had steroids only for brief periods and in low doses. These observations do not support a major importance of steroid treatment for the development of fatal a.a.d. A s pointed o u t by Werne (It%), Robinson (12) and Ornilla et al. ( 1 I), a.a.d. has to be considered as a possible risk in connection with anesthesia for surgery in RA. Physiotherapeutic manipulations may be dangerous, too.
REFERENCES I . Ball, J. & Sharp, J.: Rheumatoid arthritis of the cervical spine. In: Modern trends in rheumatology, 2nd ed. p. 117. AGS Hill, Butterworth. London 1971. 2 . Brattstrom, H . & Granholm. L.: Chirurgie der Halswirbelsaule bei Patienten mit rheumatoider Arthitis. Orthopade 2: 118, 1973. 3. - Operative treatment of atlanto-axial luxation in rheumatoid arthritis. In manuscript 1975. 4. Conlon, P. W . , Isdale, I . C. & Rose, 8 . S.:
Rheumatoid arthritis of the cervical Spine Ann. rheum. Dis. 25: 120, 1966. 5 . Davis, Jr, F. W. & Markley, H. E.: Rheumatoid arthritis with death from medullary compression. Ann. Med. 35: 451, 1951.
6. Hauge, T.: So called spontaneous cervical dislocations. Acta chir. scand., Suppl. 232: 5 , 1958. 7. Martel, E. & Abel, M. R.: Fatal atlanto-axial luxation in rheumatoid arthritis. Arthr. and Rheum. 6: 224. 1963. 8. Martel, W . & Page, J . W.: Cervical vertebral erosions and subluxations in rheumatoid arthritis and ankylosing spondylitis. Arthr. and Rheum. 3: 546. 1960. 9. Mathews, J . : Atlanto-axial subluxation in rheumatoid arthritis Ann. rheum. Dis. 28: 260. 1969. 10. Mathews, J . A,: Atlanto-axial subluxation in rheumatoid arthritis. A 5-year follow-up study. Ann. rheum. Dis. 33: 526, 1974. 1 1 . Ornilla. E., Ansell, B. M. & Swannell. A. J.: Cervical spine involvement in patients with chronic arthritis undergoing orthopedic surgery. Ann. rheum. Dis. 3 I : 364, 1972.
12. Robinson. H. S.: Rheumatoid arthritis-atlanto-axial subluxation and its clinical presentation. Canad. med. Ass. J. 94: 470, 1966. 13. Smith, P. H . , Benn. R. T. & Sharp, J.: Natural history of rheumatoid cervical luxations. Ann. rheum. Dis. 31: 431, 1972. 14. Storey. G.: Changes in the cervical spine in rheumatoid arthritis with compressions of the cord. Ann. phys. Med. 4: 216, 1958. 15. Wadia, N . H.: Myelopathy complicating congenital atlanto-axial dislocation. Brain 9 0 449, 1967. 16. Webb, F. W. S . . Hickman. J. A. & Brew, D. St. J . : Death from vertebral artery thrombosis in rheumatoid arthritis. Brit. med. J. 2: 537, 1968. 17. Werne, S . : Studies in spontaneous atlas dislocation. Acta orthop. scand.. Suppl. 23: 5, 1957. 18. - Narkosrisken vid atlasdislokation. Lakartidningen 64: 4475. 1967.
Acta med. scand. 198
