June Issue.indd

Histiocytic necrotising lymphadenitis (Kikuchi’s disease): A rare cause of cervical lymphadenopathy

Correspondence to: V. Kaushik, 26 Francis Road, Withington, Manchester M20 4XP, UK Email: viv959@yahoo.co.uk

Keywords: Histiocytic necrotising lymphadenitis, Kikuchi’s disease, lymphadenopathy, steroids
Surg J R Coll Surg Edinb Irel., 2 June 2004, 179-182

Both Head and Neck Surgeons and General Surgeons are frequently referred patients with cervical lymphadenopathy. An uncommon but important cause is histiocytic necrotising lymphadenitis. This is a benign self-limiting disease that has been confused with malignant lymphomas. Some patients may also experience distressing and debilitating symptoms which can last for months. We describe four cases to illustrate the varied clinical presentation of this disease and present new signs seen in association with it. A remarkable therapeutic response to a short course of oral corticosteroids was observed in one case

INTRODUCTION
Described over 30 years ago by Kikuchi (1972) in Japan, histiocytic necrotising lymphadenitis (HNL) was subsequently recognised and reported in the West.¹ However, there remains a distinct lack of awareness of this condition amongst both medical and surgical specialities to whom these cases may present.

Histiocytic necrotising lymphadenitis is a pathological diagnosis based on characteristic morphological and immunohistochemical features.² Clinically and histologically the condition has been mistaken for tuberculosis, lupus and lymphomas. There are reports in the literature of patients having undergone staging laparotomies, anti-tuberculous therapy and even chemotherapy.^3-6 A recent series from the UK found that it was still frequently being confused histologically with malignant lymphomas.²

We present a series of cases illustrating the characteristic features of the condition in an attempt to increase the awareness of this important differential diagnosis of cervical lymphadenopathy. We describe new signs seen in association with the disease and report the response to oral prednisolone in one case. The patients presented to our institution over a seven-year period, between 1995 and 2002.

Case 2
A 35-year old Asian female was referred with a four week history of a painful left neck lump which had not responded to a week’s course of penicillin. She also described episodes of pyrexia, rigors, and backache. Physical examination revealed two tender lymph nodes on the left side of her neck. The first measured 2 x 3cm and was located in the upper posterior triangle. The second, located in the anterior triangle at the midcervical level, was slightly smaller in size at 1 x 2 cm. Ear, nose and throat examination was unremarkable. White blood count (WBC) was 3.1 [normal: 4.0-11.0 x109/l], with a neutropenia (1.50) [normal: 2.0- 7.5 x 109/l] and lymphocytopenia (1.30) [normal: 1.5- 4.0 x109/l]. Erythrocyte sedimentation rate (ESR) was elevated at 81mm/hr [normal: 0 to 7mm/hr (female)]. Viral serology was negative for cytomegalovirus (CMV), toxoplasma, treponema, coxsackie virus and enterovirus. Epstein-Barr virus (EBV) serology was consistent with past EBV infection.

She was initially treated with a two week course of co-amoxiclav without effect. An ultrasound scan of her neck demonstrated extensive well defined lymphadenopathy on the left side extending along the carotid sheath and adjacent to sternocleidomastoid. No abnormality was seen on the right side of the neck. Fine needle aspiration (FNA) was inconclusive. Excision biopsy was performed and findings were consistent with HNL. The patient took paracetamol to relieve her discomfort. Her symptoms resolved spontaneously a few weeks later. At three months, no cervical lymphadenopathy was noted and she was consequently discharged.

Case 3
A 20 year old Caucasian female was referred with an eight-week history of a painful left neck swelling, general malaise and anorexia. In clinic, she mentioned that a further ipsilateral neck lump had developed a few days earlier. Physical examination revealed a lymph node in the left posterior triangle and one in the left anterior triangle at the mid-cervical level. Ear, nose and throat examination was unremarkable. There was no hepatosplenomegaly. WBC was 2.6 x 109/l with a neutrophil count of 1.58. ESR was 15mm/hr. The chest radiograph was normal.

Ultrasound scan of the neck demonstrated numerous large lymph nodes in the left cervical chain; the largest measured 3 x 2.8 x 1.9 cm. The right side of the neck was normal. Initial FNA was inconclusive. A repeat aspirate showed diffuse infiltration by small lymphocytes and occasional centroblastic cells. No typical Reed-Sternberg cells were seen. The pathological differential diagnosis lay between Hodgkin’s disease (as Reed-Sternberg cells can be very scarce) and severe atypical hyperplasia. The Histopathologist advised a lymph node biopsy.

The patient was listed for excision biopsy. On admission her symptoms had improved and the lymph nodes had decreased in size but it was decided to proceed for diagnostic purposes. Excision biopsy was performed and showed the appearances of HNL. As the patient was asymptomatic no medication was prescribed. At two months follow-up she was well and the lymphadenopathy had completely resolved. Her WBC returned to normal at three months. She remained asymptomatic at four months and, therefore, was discharged.

Case 4
A 34-year old Asian male was referred for excision biopsy of his cervical lymph nodes. He gave a six-week history of malaise, nocturnal low grade fevers, and myalgia. Over the preceding three weeks he had noticed cervical lymphadenopathy, experienced anorexia and lost 10 kilograms in weight. He had been treated empirically for toxoplasmosis using a combination of pyrimethamine and sulphadiazine, but this was discontinued a week later because his symptoms worsened. Subsequent viral serology (including toxoplasma serology), a chest radiograph and malarial films were unremarkable. The ESR was elevated at 50mm/hr [normal =0 to 5mm/hr (male)]; WBC was 3.3 x 109/l with a neutrophil count of 1.55 and lymphocyte count of 1.04; gamma-glutamyl transpeptidase (.-GT) was raised at 195 IU/l [normal<45 IU/l] and alanine aminotransferase (ALT) elevated to 200 IU/l [normal=5-45 IU/l].

On further questioning the patient mentioned having nocturnal vomiting, a rash affecting his body and limbs, and a one-week period of arthropathy affecting his hands. Similar episodes of malaise, lethargy and anorexia had occurred almost 11 years earlier, spanning a four-year period. They occurred approximately every six months and would last a few days. Our patient could not recall if they were associated with cervical lymphadenopathy.

Physical examination revealed dehydration and pyrexia of 38.2°C. Examination of the oropharynx showed an injected soft palate and a granular appearance of the uvula mucosa. Palpation of the neck revealed multiple bilateral 0.5 cm posterior triangle lymph nodes, with two larger 1 cm nodes palpable at the apex of the left posterior triangle. Smaller nodes were noted in both anterior triangles, the left supraclavicular fossa, right axilla and both inguinal regions. There was 1 cm hepatomegaly and the tip of the spleen was just palpable. A subtle confluent, erythematous rash affected his upper limbs, anterior chest wall and the whole of the back,and face. Nasendoscopic evaluation revealed an injected epiglottis and a diffuse midline postnasal space swelling.

The patient subsequently underwent excision biopsy of one of the posterior triangle lymph nodes and the pathological findings were consistent with HNL. In view of the severity and persisting nature of his symptoms he was treated with prednisolone (40mg once daily) for one week. He experienced significant improvement in his general condition after just two days, with complete resolution of his symptoms and signs after one week. His hepatic enzyme abnormalities returned to normal over the next month. He remains well nine months later.

DISCUSSION
Histiocytic Necrotising Lymphadenitis is an uncommon cause of cervical lymphadenopathy. It was first described independently by Kikuchi (1972) and Fujimoto et al.(1972)^1,7 However, to date, the precise aetiology remains obscure. Apoptosis is a prominent feature and thought to be related to its pathogenesis.⁸ It has been suggested that it is due to a viral or post-viral hyperimmune reaction.^9,10 Epstein-Barr virus, cytomegalovirus, varicella zoster virus, and human herpes virus six have all been implicated. Chan et al.(1989) proposed a broader pathogenetic model, stating that it was a non-specific hyperimmune reaction to a variety of infectious, chemical, physical and neoplastic agents.¹¹

There are no diagnostic laboratory tests which can confirm the diagnosis of HNL. Patients may have mild granulocytopenia, with an elevated ESR and C-reactive protein. The only reliable method of establishing the diagnosis is by excision biopsy of the enlarged lymph nodes. Histologically, the characteristic lesions of HNL are seen in the paracortex of lymph nodes. They comprise welldefined foci of necrosis with copious karyorrhectic debris (Figures 1 and 2). These areas are surrounded by a zone of large CD8 positive T-cells and macrophages. The latter commonly show C-shaped nuclei, so called “crescentic histiocytes”. Depending on the predominant cell type, HNL may be histologically subtyped as proliferative, necrotising or xanthomatous. Polymorphonuclear leukocytes and plasma cells are lacking.

Figure 1: Low power view of a lymph node showing extensive necrosis with a residual rim of lymphoid tissue (H&E)

Figure 2: High power view with necrosis and adjacent lymphoid tissue containing transformed lymphocytes and apoptotic bodies (H&E)

There are reports of HNL being successfully diagnosed by FNA of the enlarged lymph nodes. Kung et al. (1990) established the diagnosis in two cases using FNA cell blocks.¹²

This method involves examining ‘cell block’ sections prepared from tissue fragments in the aspirate. Because the architectural relationships of the different cell types are preserved in these fragments, the appearances strongly resemble excision biopsy findings, and this allows the diagnosis to be made.

Hsueh et al.(1993) diagnosed HNL in patients with typical clinical features using the more widely used FNA smear.¹³ Cytological features considered characteristic of HNL included the presence of small and large atypical lymphocytes, some reactive, phagocytic histiocytes, intense extracellular debris and the absence of neutrophils, plasma cells or multinucleated giant cells. They advised repeat nodal FNA or lymph node biopsy to establish the diagnosis in patients with typical clinical features but non-diagnostic findings in the FNA aspirate.

Typically, HNL occurs between the ages of 20 and 70 years, but most commonly in those under thirty.⁵ There are rare instances of it occurring in children.¹⁴ A definite female preponderance exists, and most patients reported with the condition have been of Asian descent.

Cervical lymphadenopathy is the commonest mode of presentation. The nodes are firm, mobile, and usually measure between 2 and 3 cm in diameter. They may be tender and can occur in isolation or as a chain. There seems to be a predilection for the posterior triangle of the neck. The supraclavicular area is also disproportionately affected but to a lesser extent.^3,15 Generalised lymphadenopathy occurs in up to 20% of cases and hepatosplenomegaly is infrequent.³

There may be associated constitutional upset including fever, nausea, vomiting and weight loss although these features tend to occur less frequently. Patients sometimes give a preceding history of an upper respiratory tract infection and it is not unusual for them to have been treated empirically with antibiotics. Cutaneous involvement has been reported but is an uncommon finding.¹⁶ The skin lesions have been described as maculopapular, drug eruption-like, rubella-like and as disseminated erythema.

Although there have been two reports of fatal cases, HNL is usually a self-limiting condition, which runs a variable course of between one to four months before resolution.^11,17,18

Single recurrences have been described.^4,19 Smith et al. (1992) reported a patient who had had multiple recurrences over an 18-year period.²⁰

Even though the condition is benign and self-limiting the accompanying symptoms can be debilitating and cause considerable ill health and distress to the patient. Patients with tender lymphadenopathy and pyrexias have traditionally been treated symptomatically using aspirin and non-steroidal antiinflammatory drugs (NSAIDs). The use of corticosteroids was initially limited to complex cases.²¹

Jang et al.(2000) suggested broadening the indications for steroid use to include patients with severe and persisting symptoms or recurrence.²¹ He reported three cases that benefited significantly from oral prednisolone. The rapid response to systemic steroids in Case 4 is consistent with Jang’s findings and supports his recommendation. We believe there may be a case for using steroids as a first-line treatment in all patients who are symptomatic as it may provide a rapid therapeutic resolution of troublesome symptoms, such as tender lymphadenopathy, pyrexia and arthropathy, that would otherwise only be treated symptomatically (using aspirin and NSAIDs) in the first instance.

The clinical features in our cases were consistent with those previously documented in the literature. The less frequent findings of hepatic dysfunction, skin lesions and arthropathy were also present in one case.²² In addition, Case 4 was noted to have a granular appearance of his uvula mucosa, an injected soft palate and epiglottis, as well as a midline postnasal space swelling. All of these signs resolved following the short course of steroids. Furthermore, his history of having experienced similar symptoms 11 years earlier suggests that his current presentation may in fact be a recurrence. Although these episodes were not investigated, nor proven histologically at the time, there nonetheless remains a strong possibility on clinical grounds. Long-term follow-up may shed further light on the natural history of this disease.

CONCLUSION
Histiocytic necrotising lymphadenitis is diagnosed by pathological assessment of the excised lymph node. Commencement of steroid therapy for patients experiencing debilitating symptoms and prolonged ill health appears to lead to a rapid resolution and recovery. A further extension of the role of steroids to include all patients who are symptomatic (in whom steroids are not contraindicated) may also be beneficial, but requires further investigation.

REFERENCES
1. Kikuchi M. Lymphadenitis showing focal reticulum cell hyperplasia with nuclear debris and phagocytes: a clinicopathological study. Acta Haematologia Japan 1972; 35: 379-80.
2. Menasce LP, Banerjee SS, Edmondson D, Harris M. Histiocytic necrotising lymphadenitis (Kikuchi-Fujimoto disease): continuing diagnostic difficulties. Histopathol 1998; 33: 248-54.
3. Turner RR, Martin J, Dorfman RF. Necrotising lymphadenitis, a study of 30 cases. Am J Surg Pathol 1983; 7: 115-23.
4. Nieman RB. Diagnosis of Kikuchi’s disease. Lancet 1990; 335: 295.
5. Dorfman RF. Histiocytic necrotising lymphadenitis of Kikuchi and Fujimoto. Arch Pathol Lab Med 1987; 111: 1026-29.
6. Papadimitriou CS, Papacharalampous NX. Histiocytic necrotising lymphadenitis without granulocytic infiltration (letter). Arch Pathol Lab Med 1985; 109: 107-8.
7. Fujimoto Y, Kojima Y, Yamaguchi K. Cervical sub-acute necrotising lymphadenitis. Naika 1972; 30: 920-27. [In Japanese]
8. Ura H, Yamada N, Torii H, Imakado S, Iozumi K, Shimada S. Histiocytic necrotising lymphadenitis (Kikuchi’s disease): The necrotic appearance of the lymph node cells is caused by apoptosis. J Dermatol 1999; 26: 385-89.
9. Nikanne E, Ruoppi P, Vornanen M. Kikuchi’s disease: report of three cases and an overview. Laryngoscope 1997; 107: 273-76.
10. Imamura M, Ueno H, Matsuura A, Kamiya H, Suzuki T, Kikuchi K, Onoe T. An ultrastructural study of subacute necrotising lymphadenitis. Am J Pathol 1982; 107: 292-99.
11. Chan JKC, Wong K, Ng C. A fatal case of multicentric Kikuchi’s histiocytic necrotising lymphadenitis. Cancer 1989; 63: 1856-62.
12. Kung ITM, Ng WF, Yuen RWS, Chan JKC. Kikuchi’s Histiocytic Necrotising Lymphadenitis: diagnosis by fine needle aspiration. Acta Cytologica 1990; 34: 323-28.
13. Hseuh EJ, Ko WS, Hwang WS, Yam LT. Fine needle aspiration of histiocytic necrotising lymphadenitis (Kikuchi’s disease). Diagnostic Cytopathology 1993; 9: 448-52.
14. Emir S, Gogus S, Guler E, Buyukpamukcu M. Kikuchi-Fujimoto disease (Histiocytic necrotising lymphadenitis) confused with lymphoma in a child. Med Pediatr Oncol 2001; 37: 546-48.
15. Garcia CE, Girdhar-Gopal HV, Dorfman DM. Kikuchi-Fujimoto disease of the neck update. Ann Otol Rhinol Laryngol 1993; 102: 11-15.
16. Seno A, Torigoe R, Shimoe K et al. Kikuch’s disease (histiocytic necrotising lymphadenitis) with cutaneous involvement. J Am Acad Dermatol 1994; 30: 504-6.
17. Lin SH, Ko WS, Lee HS, Hwang WS. Kikuchi’s disease associated with lupus-like syndrome- a fatal case [Letter]. J Rheumatol 1992; 19: 1995-96.
18. Hoyt DJ, Fisher SR. Kikuchi’s disease causing cervical lymphadenopathy. Otolaryngol Head Neck Surg 1990; 102: 755-58.
19. Dorfman RF, Berry GJ. Kikuchi’s histiocytic necrotising lymphadenitis: an analysis of 108 cases with emphasis on differential diagnosis. Sem Diag Pathol 1988; 5: 329-45.
20. Smith KGC, Becker GJ, Busmanis I. Recurrent Kikuchi’s disease [Letter]. Lancet 1992; 340: 124.
21. Jang YJ, Park KH, Seok HJ. Management of Kikuchi’s disease using glucocorticoid. J Laryngo Otol 2000; 114: 709-11.
22. Bailey EM, Klein NC, Cunha BA. Kikuchi’s disease with liver dysfunction presenting as fever of unknown origin. Lancet 1989; ii: 986.

Re: Bier’s block using prilocaine: Safe cheap and well tolerated. Surg J R Coll Edinb Irel 2003; 1(5):283-5.

I was pleased to read the above article on the use of prilocaine Bier’s block for Colles’ fracture manipulation. I would like to offer some observations on the methods, reporting and conclusions of this study.

Most researchers inevitably have to accept an element of taking things on trust but this is especially true of a retrospective review of hospital records. How sure can we be that all procedures were rigorously logged and that all adverse events were correctly and systematically recorded throughout the 15 years covered by the retrospective review? Therefore, I would have liked to see an account of the quality checks undertaken by the authors as well as an assessment and, if appropriate and possible, a quantitative estimate of the problem of missing and incomplete data. There is also potential for errors, such as the overlooking of relevant information, in the process of reviewing operation logbooks. Such errors can be reduced by using systematic methods of detection, including quality checks. I suggest that further details of the methods of detection used would also be informative.

Pickering and Hunter refer to a now dated survey of anaesthetic practice for Colles’ fracture in the UK.¹ A subsequent published survey conducted in 1994 of anaesthetic practice for these fractures was based on responses from 86 accident and emergency departments in the UK.2 This revealed that Bier’s block was given to one third of patients, haematoma block to another third, general anaesthesia to 24%, and intravenous sedation to 7%.² I do not myself know whether these results reflect current practice. Whatever, it is likely that the authors’ assertion that there is “still no consensus on the best technique” holds. This also reflects the currently insufficient evidence for determining the relative effectiveness of different methods of anaesthesia.³ There is, however, some indication that haematoma block provides poorer analgesia than Bier’s block, and can compromise reduction.³

The authors rightly point out some limitations of their questionnaire-based study, which nonetheless provides some useful insights on what aspects of the procedure cause distress to patients. However, the authors do not provide any evidence in support of their claims that their “Colles fracture results are consistent with those throughout the country”. Nor does their work provide a basis for drawing conclusions on the relative effectiveness of Bier’s block compared with other anaesthetic techniques.

REFERENCES
1. Hunter JB, Scott MJL, Harries SA. Methods of anaesthesia used for reduction of Colles’ fractures. BMJ 1989; 299:1316-7.
2. Kendall JM, Allen PE, McCabe SE. A tide of change in the management of an old fracture? J Accid Emerg Med 1995; 12(3): 187-8.
3. Handoll HHG, Madhok R, Dodds C. Anaesthesia for treating distal radial fracture in adults (Cochrane Review). In: The Cochrane Library. Chichester, UK: John Wiley & Sons, Ltd, 2003(4).

Dr H. Handoll
University Department of Orthopaedic
Surgery, Royal Infirmary of Edinburgh, Little France, Old Dalkeith Road, Edinburgh, EH16 4SU

Re: Bier’s block using prilocaine: safe,cheap and well tolerated. Surg J R Coll Edinb Irel 2003; 1: 283-285.

I read with interest the above article by Pickering and Hunter. The authors referred to Bier’s block as cheap (title and discussion) and inexpensive (conclusion). However, the issue of the cost of Bier’s block was not addressed at all in the retrospective study performed by the authors. The authors made comments on costs by citing the article of Funk (1997). Funk’s study included general anaesthesia and haematoma block (with and without sedation), but not Bier’s block. The authors also cited the article by Cobb et al. (1985) which addressed various issues but no comment was made on the cost.

Bier’s block may well be cheap and inexpensive. However, based on the ‘evidence’ provided in the article, could the authors really draw any conclusion on the cost of Bier’s block?

REFERENCES
1. Funk L. A prospective trial to compare three anaesthetic techniques used for the reduction of fractures of the distal radius. Injury 1997; 28: 209-212.
2. Cobb AG, Houghton GR. Local anaesthetic infiltration versus Bier’s block for Colles’ fracture. BMJ 1985; 291: 1683-1684.

We thank Mr Debnath for his comments. He is absolutely right to point out that we have not specifically quoted costs for the common treatment options. However, we felt that anyone reading this paper would appreciate that the cost of hospital admission, a general anaesthetic and overnight stay far outweighs the cost of a bottle of local anaesthetic and discharge several hours later from fracture clinic.

We thank Dr Handoll for their comments. The surgical log books kept in the fracture manipulation room are the most reliable source of information on patients undergoing simple fracture manipulations in our institution. All details are rigorously recorded before patients receive treatment, indeed treatment would not occur unless the log book was present. We were looking for rare, serious events, such as arrythmia or anaphylaxis, occurring during administration of the local anaesthetic. Although possible, it is highly unlikely that a patient could have a life threatening complication without this being noted in the logbook, either at the time of the adverse event or retrospectively.

We feel that Dr Handoll has rather missed the point of this article. Even one patient having a cardiac arrhythmia or anaphylaxis in fifteen years would be considered a significant event and certainly one remembered by more senior members of staff in our institution. Anecdotally, we believed this to be a safe technique which was easily confirmed by reference to the serious adverse events column in the fracture room logbook. In order to give an exact number of procedures and accurate patient demographics all patients had to be counted and their ages recorded. There is the possibility that occasional patients may have been missed despite recounting. Ultimately, this will not have affected the conclusion of this article.

In a recent study at our institution, Earnshaw et al.(2002) found that 87% of 225 patients with a displaced Colles fracture had an acceptable fracture reduction using a Biers block.1 An acceptable reduction was considered to be a dorsal tilt of less than 10º and radial shortening of less than 5mm. These results are in keeping with another recent study by Young et al.(2003)2 In 50 patients with a displaced radial fracture treated in plaster after manipulation, they found that the 25th to 75th percentile range was 2º to 8º for dorsal angulation and 2mm to 5mm for radial shortening.

REFERENCES
1. Earnshaw SA, Aladin A, Surendran S, Moran CG. Closed reduction of Colles fractures: comparison of manual manipulation and finger-trap traction: a prospective, randomised study. Journal of Bone and Joint Surgery 2002; 84-A(3): 354-8.
2. Young CF, Nanu AM, Checketts RG. Seven-year outcome following Colles’ type distal radial fracture. A comparison of two treatment methods. Journal of Hand Surgery 2003; 28(5):422-6.

Mr S.A.W. Pickering and Mr J.B. Hunter
Department of Trauma and Orthopaedic Surgery, Queen’s Medical Centre,
Nottingham, NG7 2UH, UK