Embedded Files
頭痛不一定只是壓力大或肩頸緊。2024 年 Emergency Medicine Journal 的 HEAD 與 HEAD-Colombia 研究分析 5293 位急診頭痛患者，發現新的神經學缺損、癌症病史、50 歲以上與近期頭部外傷，是較重要的嚴重頭痛紅旗。本文整理頭痛什麼時候該去急診、哪些情況不能只靠止痛藥硬撐。
頭痛什麼時候該去急診？研究發現：真正危險的不是痛幾分，而是這些紅旗症狀
頭痛什麼時候該去急診？研究發現：真正危險的不是痛幾分，而是這些紅旗症狀
很多人都有頭痛經驗。
睡不好會頭痛，壓力大會頭痛，肩頸太緊會頭痛，感冒會頭痛，月經前後也可能頭痛。也因為頭痛太常見，很多人第一時間會選擇忍一忍、按一按、喝咖啡，或直接吃止痛藥。
但問題是：有些頭痛真的不能忍。
不是因為它一定很痛，而是因為它可能不是一般偏頭痛或緊繃型頭痛，而是其他疾病發出的警訊。例如腦出血、蜘蛛膜下腔出血、腦膜炎、腦腫瘤、中風、顱內壓升高、血管剝離、顳動脈炎，甚至懷孕相關高血壓或子癲前症，都可能以「頭痛」作為表現。
2024 年發表於 Emergency Medicine Journal 的研究 Predictive performance of the common red flags in emergency department headache patients: a HEAD and HEAD-Colombia study，分析 HEAD 與 HEAD-Colombia 兩個研究中的急診頭痛患者資料，總共納入 5293 位成人急診頭痛患者。研究者想回答一個很實際的問題：我們平常說的頭痛紅旗症狀，到底哪些真的比較能預測嚴重次發性頭痛？
研究結果發現，5293 位急診頭痛患者中，有 323 位，也就是 6.1%，最後被確認有嚴重次發性頭痛。這些嚴重病因包括顱內腫瘤、顱內出血、腦膜炎、蜘蛛膜下腔出血、中風、特發性顱內高壓、顳動脈炎、水腦、腦炎、血管剝離、腦膿瘍、高血壓危象、懷孕相關高血壓等。
這個數字很重要。因為它告訴我們兩件事。
第一，大多數急診頭痛最後不是嚴重疾病。
第二，少數嚴重頭痛真的存在，而且不能被漏掉。
所以，頭痛衛教的重點不是讓大家恐慌，而是讓大家知道：什麼時候可以觀察，什麼時候不能硬撐。
什麼是「次發性頭痛」？為什麼它比一般頭痛更需要小心？
什麼是「次發性頭痛」？為什麼它比一般頭痛更需要小心？
頭痛大致可以分成兩大類：原發性頭痛與次發性頭痛。
原發性頭痛，是指頭痛本身就是主要疾病。常見像偏頭痛、緊繃型頭痛、叢發性頭痛。這些頭痛雖然很痛、很影響生活，但不是由腦出血、感染、腫瘤等其他疾病直接引起。
次發性頭痛，則是指頭痛是其他疾病造成的結果。也就是說，頭痛只是表面訊號，背後可能藏著腦血管問題、感染、顱壓異常、腫瘤、外傷或全身性疾病。
這篇研究把「嚴重次發性頭痛」定義為一系列需要高度警覺的診斷，包括蜘蛛膜下腔出血、顱內出血、腦膜炎、腦炎、腦膿瘍、顱內腫瘤、水腦、血管剝離、中風或短暫性腦缺血、 hypertensive crisis、懷孕相關高血壓或子癲前症、顳動脈炎、特發性顱內高壓，以及腦室腹腔分流管併發症。
這些疾病有些需要急診處置，有些需要神經科、感染科、神經外科或影像檢查介入。也因此，面對頭痛時，最重要的不是先問「哪一種止痛藥比較強」，而是先判斷：這次頭痛有沒有可能不是普通頭痛？
頭痛紅旗是什麼？不是診斷，而是警報系統
頭痛紅旗是什麼？不是診斷，而是警報系統
所謂「頭痛紅旗」，就是一些會讓醫師提高警覺的線索。
它的作用不是直接診斷某一種疾病，而是提醒我們：這個頭痛不一定能當作一般偏頭痛或緊繃型頭痛處理，可能需要進一步檢查。
常見紅旗包含發燒、癌症病史、新的神經學缺損、突然爆炸性頭痛、50 歲以上新發頭痛、運動或性行為誘發頭痛、視乳突水腫、懷孕或產後、近期頭部外傷、免疫系統問題等。這篇研究也針對這 10 項常見 SNNOOP10 紅旗條件進行分析。
但這篇文獻最重要的提醒是：紅旗不是完美工具。
研究發現，這些紅旗合併使用時，敏感度很高，代表它們有助於抓出較高風險患者；但特異性非常低，代表很多有紅旗的人，最後不一定真的有嚴重疾病。整體來看，10 項紅旗合併使用時，敏感度為 96.5%，但特異性只有 5.1%，陽性預測值只有 9.3%，陰性預測值為 93.5%。
白話來說就是：
有紅旗，不代表一定有大病。
但有紅旗，代表不能隨便忽略。
這就像大樓的火警警報器。警報響，不代表一定真的失火；有時候可能是誤報。但警報響了，你不能只把聲音關掉，還是要確認是不是有真正的危險。
最有預測力的紅旗一：新的神經學缺損
最有預測力的紅旗一：新的神經學缺損
這篇研究發現，「新的神經學缺損」是嚴重次發性頭痛的重要預測因子。在整體分析中，新的神經學缺損與嚴重次發性頭痛有強烈關聯，調整後勝算比為 6.63。
什麼叫新的神經學缺損？
簡單說，就是頭痛不只是痛，還合併神經功能異常。例如突然單側手腳無力、臉歪、說話不清、語言表達困難、走路不穩、視野缺損、意識混亂、反應變差、抽搐，或昏睡叫不太醒。
這類情況不能只當作一般頭痛。
有些偏頭痛患者也可能出現視覺預兆、手麻、說話卡住，所以不是所有神經症狀都一定是中風。但如果這些症狀是第一次出現，或和過去偏頭痛不同，或持續不退，或越來越嚴重，就不應該在家觀察。
尤其是頭痛合併說話不清、單側無力、意識混亂，這不是單純疼痛問題，而是大腦功能正在出現異常訊號。
如果用白話比喻，一般頭痛像是警報器響；頭痛合併手腳無力、說話不清，則像是警報響的同時，電梯停了、燈也開始閃。這時候不能只想把警報關掉，而是要立刻確認整棟大樓是不是出問題。
最有預測力的紅旗二：癌症病史
最有預測力的紅旗二：癌症病史
研究中另一個強烈預測因子是「癌症病史」。在整體分析中，癌症病史的調整後勝算比為 7.82，是所有紅旗中關聯最強的一項。
這並不是說有癌症病史的人一頭痛就一定是腦轉移。癌症患者也可能有偏頭痛、緊繃型頭痛、睡眠不足或肩頸緊造成的頭痛。
但如果曾經有癌症病史，又出現新的頭痛、頭痛型態改變、越來越嚴重，或伴隨嘔吐、癲癇、視力變化、意識改變、肢體無力，就需要更小心。
因為這類頭痛背後可能需要排除顱內腫瘤、轉移、顱壓升高、感染、血栓，或治療相關併發症。
這也提醒我們，判斷頭痛不能只看「痛不痛」，還要看「這個人的背景風險」。同樣是頭痛，一位年輕健康、過去就有典型偏頭痛的人，和一位有癌症病史、最近突然出現新頭痛的人，風險完全不同。
最有預測力的紅旗三：50 歲以上
最有預測力的紅旗三：50 歲以上
這篇研究也發現，年齡大於 50 歲與嚴重次發性頭痛有顯著關聯，調整後勝算比為 2.00。
為什麼 50 歲以後的新頭痛要特別注意？
因為偏頭痛、緊繃型頭痛雖然可以延續到中年以後，但多數原發性頭痛通常比較早就開始出現。如果一個人過去不太頭痛，到了 50 歲後才突然出現新的頭痛型態，臨床上就會更重視是否有其他病因。
例如顳動脈炎、腦血管疾病、顱內腫瘤、顱壓異常、感染、眼壓問題、藥物影響，或全身性發炎疾病，都可能在中高齡族群以頭痛表現。
對 40 到 55 歲的商業白領族群來說，這個觀念特別重要。很多人會把頭痛歸因於工作壓力、睡不好、肩頸緊。但如果進入中年後，頭痛型態開始改變，或以前不太頭痛、現在突然常常頭痛，就不建議只用「最近太累」來解釋。
最有預測力的紅旗四：近期頭部外傷
最有預測力的紅旗四：近期頭部外傷
研究也發現，近期頭部外傷，特別是頭痛前 3 到 7 天內的頭部外傷，與嚴重次發性頭痛有關，調整後勝算比為 2.67。
很多人撞到頭之後，如果當下沒有昏倒、沒有流血，就會覺得應該沒事。但有些顱內出血、硬腦膜下出血或腦震盪後症候群，並不一定在撞到當下最明顯，而可能在幾天後才出現頭痛、嘔吐、嗜睡、反應變慢、走路不穩。
尤其年長者、正在服用抗凝血藥或抗血小板藥物的人，更要小心。
如果撞到頭後幾天內頭痛越來越明顯，或合併嘔吐、嗜睡、意識混亂、單側無力、走路不穩，就不要只在家休息，應盡快就醫評估。
令人意外的發現：突然爆痛，不是單獨判斷的全部
令人意外的發現：突然爆痛，不是單獨判斷的全部
很多人一聽到「突然爆炸性頭痛」，第一個想到就是蜘蛛膜下腔出血。臨床上，雷擊樣頭痛確實是非常重要的危險訊號。
但這篇研究有一個值得細讀的結果：在調整其他預測因子後，「突然發作」本身並沒有和嚴重次發性頭痛呈現獨立顯著關聯。整體分析中，突然發作頭痛的調整後勝算比為 1.43，信賴區間跨過 1；在沒有神經學異常的次族群中，突然發作頭痛同樣不是獨立顯著預測因子。
這個結果不能被誤解成「突然爆痛不用看醫生」。
比較正確的理解是：
在急診真實世界裡，單一紅旗的預測能力有限。突然爆痛仍然重要，但它不能脫離其他線索單獨判斷。
如果突然爆痛合併頸部僵硬、意識改變、神經缺損、嘔吐、血壓異常、過去沒有類似經驗，仍然應該盡快就醫。這篇研究提醒我們的是：臨床判斷不能只靠一個關鍵字，而要整合整個病人的狀態。
頭痛到吐，是不是一定很危險？
頭痛到吐，是不是一定很危險？
很多患者會問：「頭痛到吐是不是中風？是不是腦壓高？」
答案是：不一定，但要看情境。
偏頭痛本來就可能伴隨噁心、嘔吐、怕光、怕聲。因此，如果你過去有明確偏頭痛診斷，每次發作模式都很類似，頭痛到吐不一定代表急症。
但如果這是第一次頭痛到吐，或這次頭痛和過去不同，或伴隨突然爆痛、神經學缺損、意識混亂、發燒、頸部僵硬、近期頭部外傷、癌症病史、50 歲後新發頭痛，就不能只當作一般偏頭痛。
這裡要強調：
嘔吐本身不是唯一判斷標準。真正重要的是整個頭痛型態和伴隨症狀。
如果頭痛像過去熟悉的偏頭痛，處理方式可以依照醫師先前建議。
如果頭痛是陌生的、突然的、越來越嚴重的，或合併其他紅旗，就應該就醫。
發燒頭痛要注意，但也要看有沒有神經症狀
發燒頭痛要注意，但也要看有沒有神經症狀
在整體分析中，發燒並不是獨立顯著預測因子；但在「沒有神經學缺損、沒有意識改變」的次族群分析中，發燒與嚴重次發性頭痛有關，調整後勝算比為 3.27。
這個結果很實用。
如果一個人頭痛合併明顯神經症狀，醫師本來就會高度警覺。真正困難的是：看起來沒有神經學異常、意識也清楚，但頭痛合併發燒，這時候仍然要小心感染性原因，例如腦膜炎、腦炎，或其他全身感染造成的頭痛。
一般感冒也可能頭痛，但如果頭痛合併高燒、頸部僵硬、畏光嚴重、意識混亂、皮疹，或免疫力低下，就不應該只當作普通感冒。
視乳突水腫：重要，但急診現場常常沒有檢查到
視乳突水腫：重要，但急診現場常常沒有檢查到
這篇研究也提到一個臨床現實：視乳突水腫可能與嚴重病因有關，但因為 88.7% 的患者沒有接受眼底鏡檢查，因此無法把視乳突水腫納入多變項迴歸分析。研究者也指出，眼底檢查在急診現場執行率低，而且準確性也可能有限，因此未來需要更客觀、可靠、容易取得的方式來評估眼底。
視乳突水腫常和顱內壓升高有關。若患者頭痛合併視力模糊、短暫視力黑掉、耳鳴、早晨頭痛、噁心嘔吐，或醫師檢查懷疑眼底異常，就需要進一步評估顱壓相關問題。
這也提醒我們：有些危險頭痛，不一定只靠問診就能完全判斷，必要時仍需要理學檢查、影像檢查或其他檢查協助。
為什麼紅旗有用，卻不能單獨決定要不要檢查？
為什麼紅旗有用，卻不能單獨決定要不要檢查？
這篇研究的結論非常謹慎。它指出，SNNOOP10 紅旗的高敏感度表示它們可以幫助辨識較高風險的嚴重次發性頭痛患者；但是敏感度還沒有高到可以只因為「沒有紅旗」就決定不做進一步檢查。加上特異性低，若過度依賴紅旗，可能會增加 CT 掃描比例。
換句話說，紅旗是工具，不是答案。
醫師真正做的是整合判斷：
患者年齡、病史、頭痛型態、神經學檢查、生命徵象、發作速度、用藥狀況、懷孕狀態、外傷史、免疫狀態、癌症病史，以及這次頭痛是否和過去不同。
對民眾來說，最重要的不是自己背完整套 SNNOOP10，而是把關鍵資訊講清楚。
例如就醫時可以告訴醫師：
這次頭痛什麼時候開始？
是慢慢痛，還是突然到最痛？
以前有沒有類似頭痛？
這次和以前有沒有不同？
有沒有手腳無力、說話不清、視力改變？
有沒有發燒、頸部僵硬？
最近有沒有撞到頭？
有沒有癌症病史、免疫疾病、懷孕或產後？
有沒有正在吃抗凝血藥或長期止痛藥？
這些資訊，比單純說「我頭很痛」更能幫助醫師判斷。
中醫怎麼看頭痛紅旗？先排急症，再談辨證
中醫怎麼看頭痛紅旗？先排急症，再談辨證
從中醫角度來看，頭痛當然可以辨證。
有些人是風寒頭痛，怕冷、頸項僵硬、遇風加重。
有些人是風熱頭痛，發熱、喉嚨痛、口乾、頭脹。
有些人是肝陽上亢，壓力大、易怒、眼脹、血壓波動。
有些人是痰濕上擾，頭重如裹、腦袋昏沉、噁心、胃脹。
有些人是瘀血阻絡，痛點固定、刺痛、久痛不癒。
有些人是氣血不足，疲倦、頭暈、工作一久就痛。
但在現代臨床中，頭痛的第一步不是急著辨證開藥，而是先判斷有沒有危險訊號。
如果是長期反覆的偏頭痛、緊繃型頭痛、肩頸壓力型頭痛，中醫可以從調肝、化痰、活血、補氣血、改善睡眠與自律神經穩定度切入。
但如果是頭痛合併新的神經學缺損、癌症病史、新發於 50 歲後、近期頭部外傷後惡化、發燒頸僵、意識混亂、視力變化，這些情況應先排除急症，再談體質調理。
這不是中醫和西醫誰比較重要，而是醫療安全的順序：
先分急緩，再談寒熱虛實。
先排危險，再做長期調理。
結論：頭痛不是看痛幾分，而是看有沒有「不一樣」
結論：頭痛不是看痛幾分，而是看有沒有「不一樣」
這篇 HEAD 與 HEAD-Colombia 研究給我們最重要的提醒是：在急診頭痛患者中，真正嚴重的次發性頭痛大約佔 6.1%，比例不高，但不能漏掉。比較有預測力的紅旗包括新的神經學缺損、癌症病史、50 歲以上，以及近期頭部外傷；若沒有神經學異常，發燒也需要特別注意。
但紅旗不是完美答案。它們敏感度高、特異性低，有助於提高警覺，卻不能取代醫師的整體判斷。
所以，民眾最該記住的不是「頭痛幾分才要就醫」，而是：
這次頭痛有沒有和平常不一樣？
有沒有合併神經功能異常？
有沒有癌症病史？
是不是 50 歲後新出現？
最近有沒有撞到頭？
有沒有發燒、頸部僵硬或意識改變？
頭痛不是忍耐力測驗。
有些頭痛只是壓力警報。
但有些頭痛，是身體提醒你：這次真的不能硬撐。
Predictive performance of the common red flags in emergency department headache patients: a HEAD and HEAD- Colombia study
Kevin Chu,1 Anne- Maree Kelly ,2,3 Win Sen Kuan,4 Frances B Kinnear,5,6 Gerben Keijzers,7 Daniel Horner ,8 Said Laribi,9 Alejandro Cardozo,10 Mehmet Akif Karamercan,11 Sharon Klim,2 Tissa Wijeratne,12 Sinan Kamona,13,14 Colin A Graham,15 Richard Body ,16,17 Tom Roberts ,18,19 HEAD and HEAD- Colombia study groups
To cite: Chu K, Kelly A- M, Kuan WS, et al. Emerg Med J Epub ahead of print: [please include Day Month Year]. doi:10.1136/ emermed-2023-213461
Handling editor Carl Marincowitz
► Additional supplemental material is published online only. To view, please visit the journal online (https:// doi. org/ 10. 1136/ emermed- 2023- 213461).
For numbered affiliations see end of article.
Correspondence to Professor Anne- Maree Kelly, JECEMR, Western Health, Footscray, VIC 3011, Australia; anne- maree. kelly@ wh. org. au
Received 27 June 2023 Accepted 25 March 2024
© Author(s) (or their employer(s)) 2024. No commercial re- use. See rights and permissions. Published by BMJ.
ABSTRACT Objectives Only a small proportion of patients presenting to an ED with headache have a serious cause. The SNNOOP10 criteria, which incorporates red and orange flags for serious causes, has been proposed but not well studied. This project aims to compare the proportion of patients with 10 commonly accepted red flag criteria (singly and in combination) between patients with and without a diagnosis of serious secondary headache in a large, multinational cohort of ED patients presenting with headache. Methods Secondary analysis of data obtained in the HEAD and HEAD- Colombia studies. The outcome of interest was serious secondary headache. The predictive performance of 10 red flag criteria from the SNNOOP10 criteria list was estimated individually and in combination. Results 5293 patients were included, of whom 6.1% (95% CI 5.5% to 6.8%) had a defined serious cause identified. New neurological deficit, history of neoplasm, older age (>50 years) and recent head trauma (2–7 days prior) were independent predictors of a serious secondary headache diagnosis. After adjusting for other predictors, sudden onset, onset during exertion, pregnancy and immune suppression were not associated with a serious headache diagnosis. The combined sensitivity of the red flag criteria overall was 96.5% (95% CI 93.2% to 98.3%) but specificity was low, 5.1% (95% CI 4.3% to 6.0%). Positive predictive value was 9.3% (95% CI 8.2% to 10.5%) with negative predictive value of 93.5% (95% CI 87.6% to 96.8%). Conclusion The sensitivity and specificity of the red flag criteria in this study were lower than previously reported. Regarding clinical practice, this suggests that red flag criteria may be useful to identify patients at higher risk of a serious secondary headache cause, but their low specificity could result in increased rates of CT scanning. Trial registration number ANZCTR376695.
INTRODUCTION Only a small proportion of patients presenting to an ED with headache have a serious cause for their headache identified after assessment and investi-gation—about 7% in recent studies.1 2 Some with serious pathology are more obvious, such as those presenting with altered conscious state and/or
new neurological features (other than headache). The challenge for ED clinicians is to decide which patients without obvious neurological findings require additional investigation to rule out a serious secondary headache cause.
The ‘red flags’ approach has been promoted and is included in highly respected guidelines.3 Some years ago, the American Headache Society proposed the SNOOP4 criteria (Systemic signs, Neurological features, Onset sudden, Older age, Progression, Papilloedema, Positional or Pregnancy).4 More recently, Do et al expanded the list to include 15
WHAT IS ALREADY KNOWN ON THIS TOPIC ⇒A small proportion of patients who present to EDs with headache (about 7%) have serious pathology diagnosed. A challenge for ED clinicians is determining which patients (especially those with a normal neurological examination) require further investigation. So- called red flag criteria have been proposed to assist in identification of patients who are at higher risk of serious pathology and to inform decision- making about investigation. There has been limited validation of these criteria in the ED setting with mixed results.
WHAT THIS STUDY ADDS ⇒Sensitivity of the SNNOOP10 criteria as a group was high, but specificity was very low. The results challenge the predictive utility of some of the red flags. Funduscopy may be a predictor but was rarely performed.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
⇒Regarding clinical practice, this study suggests that red flag criteria are useful to identify patients at higher risk of a serious secondary headache cause, but their absence alone should not be used to determine whether further investigation is required. The low rate of funduscopy and its reported inaccuracy suggest that new and more accurate ways of examining the optic fundus may be needed.
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
by copyright.
on A pril 30, 2024 at B
en G urion U
ni M A
LM A
D C
onsortia. P rotected
http://em j.bm
j.com /
erg M ed J: first published as 10.1136/em
erm ed-2023-213461 on 24 A
pril 2024. D ow
nloaded from
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
red and orange flags for secondary headache (the so- called SNNOOP10 criteria) (figure 1).5 These have had limited eval-uation in the ED headache population and come from studies with different methodologies and had small sample sizes (fewer than 350 patients in total).6 7 One of those small studies of 100 patients reported sensitivity of SNNOOP10 list of 100% (95% CI 90.2% to 100%).7
The HEAD and HEAD- Colombia studies are multinational studies of patients presenting with headache to ED.1 2 Their data provide an opportunity to evaluate the common red flags for serious headache in a large, real- world ED population.
The main aim of this study was to explore the association between, and predictive value for serious secondary head-ache of, 10 commonly accepted red flag criteria (singly and in combination) using a large, multinational cohort of ED patients presenting with headache. A planned secondary objective was
to explore this association in the subgroup of patients who did not present with altered conscious state, new confusion or new neurological signs on examination. The rationale for this subgroup is that, for people presenting with new neurological features (in addition to headache), it is usually clear that they require investigation. The group with normal neurology poses the main challenge for ED clinicians with respect to diagnostic decision- making, including selection of investigations.
METHODS Study design and setting This was an unplanned analysis of data collected in the HEAD and HEAD- Colombia studies.1 2 Both were observational studies of adult ED patients with headache. The HEAD study was a multinational study conducted in Australia, New Zealand, Singa-pore, Hong Kong, UK, France, Belgium, Romania, Turkey and Israel. The HEAD Columbia study was undertaken in Colombia. There were 69 healthcare facilities across 11 countries (online supplemental table 1). Their methodology has been published previously.1 2
Data sources The HEAD and HEAD- Colombia studies used the same protocol and collected the same data with minor variation due to availability of some medications in some countries. (supple-mental file 1) Data were collected in 2019 and 2021 for the two studies, respectively. One of the HEAD study authors (A- MK) was involved in the HEAD Columbia study but investigators from the latter were not involved in the planning of the original HEAD study.
We considered that combining the data from these studies was valid because they used the same methodology and the same data collection tool and covered approximately the same time period. Moreover, we analysed pooled patient- level data from the two studies. A meta- analysis, on the other hand, typically analyses study- level results from multiple studies. Our data were thus in keeping with a single multicentre study, and hence were analysed as such and not as a meta- analysis.
Data collected included data on medical history and medi-cations, headache features, examination findings, patterns
Figure 1 SNNOOP10 criteria.5
Table 1 Population characteristics stratified by serious secondary headaches
HEAD and HEAD- Colombia studies HEAD study HEAD- Colombia study
Non- serious secondary headache
Serious secondary headache Total
Non- serious secondary headache
Serious secondary headache Total
Non- serious secondary headache
Serious secondary headache Total
N=4970 (93.9%)
N=323 (6.1%) N=5293
N=4276 (94.3%)
N=260 (5.7%) N=4536
N=694 (91.7%)
N=63 (8.3%) N=757
Age
Median (IQR), years 40 (29–54) 53 (35–69) 40 (29–55) 40 (29–54) 54 (36–71) 41 (29–55) 38 (28–50) 46 (31–62) 39 (28–51)
Female, n (%) 3309 (66.6) 179 (55.4) 3488 (65.9) 2767 (64.7) 140 (53.9) 2907 (64.1) 542 (78.1) 39 (61.9) 581 (76.8)
Referred by, n (%)
Self 4162 (83.7) 238 (73.7) 4400 (83.1) 3554 (83.1) 194 (74.6) 3748 (82.6) 608 (87.6) 44 (69.8) 652 (86.1)
Doctor 808 (16.3) 85 (26.3) 893 (16.9) 722 (16.9) 66 (25.4) 788 (17.4) 86 (12.4) 19 (30.2) 105 (13.9)
Mode of arrival, n (%)
Non- ambulance 4209 (84.7) 226 (70.0) 4435 (83.8) 3570 (83.5) 175 (67.3) 3745 (82.6) 639 (92.1) 51 (81.0) 690 (91.2)
Ambulance 761 (15.3) 97 (30.0) 858 (16.2) 706 (16.5) 85 (32.7) 791 (16.2) 55 (7.9) 12 (19.1) 67 (8.9)
Triage category, n (%)
Immediate 53 (1.1) 35 (10.8) 88 (1.7) 45 (1.1) 32 (12.3) 77 (1.7) 8 (1.2) 3 (4.8) 11 (1.5)
Urgent 2793 (56.2) 241 (74.6) 3034 (57.3) 2112 (49.4) 182 (70.0) 2294 (50.6) 681 (98.1) 59 (93.7) 740 (99.2)
Non- urgent 2124 (42.7) 47 (14.6) 2171 (41.0) 2119 (49.6) 46 (17.7) 2165 (47.7) 5 (0.7) 1 (1.6) 6 (0.8)
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
of investigation and final ED and hospital diagnosis in adult patients presenting to ED with acute non- traumatic headache (absence of head trauma within 48 hours of ED presentation).
Outcome The primary outcome was serious secondary headache defined as any of the following: subarachnoid haemorrhage (SAH), intracranial haemorrhage (ICH), meningitis, encephalitis, cere-bral abscess, intracranial neoplasm, hydrocephalus, vascular dissection, stroke/transient ischaemic attack, hypertensive crisis, pregnancy- related hypertension/eclampsia, temporal arteritis, idiopathic intracranial hypertension (IIH) and ventriculoperito-neal (VP) shunt complications. Final ED diagnosis was used for patients discharged from ED and the final hospital diagnosis was used for patients admitted to hospital.
Overall cohort and subgroup analyses The outcome was analysed in the overall cohort and in the subgroup without neurological findings. In the overall cohort, 10 commonly accepted SNNOOP10 red flag criteria were exam-ined. The 10 criteria were fever (>38°C), history of neoplasm, neurologicaldeficit(newfocalneurologicalsignsorGCS≤12),sudden- onset headache, age >50 years, headache precipitated by exertion including sexual activity, papilloedema, pregnancy or puerperium, recent trauma (between 3 and 7 days previously) and pathology of the immune system.5 History of neoplasm included cerebral or non- cerebral malignant neoplasm. Sudden onset was described as peaking instantly or almost instantly. Pathology of immune system was defined as chemotherapy, immunosuppressant medication, HIV, intravenous drug user or systemic lupus erythematosus. It should be noted that the HEAD studies did not collect data on the other five SNNOOP10 criteria (positional nature, painful eye with autonomic features, headache pattern change or new onset, painkiller overuse and progressive headache with atypical features) because its design preceded publication of the SNNOOP10 list.
The association between serious secondary headaches and SNNOOP10 red flags was sought using a multivariate binary logistic regression analysis of patients with non- missing data for all red flags. Specifically, the outcome was serious secondary
headaches. The predictor variables were fever, neoplasm, neuro-logical deficit, sudden- onset headache, age >50 years, exertion or sexual activity, pregnancy or puerperium, head trauma and pathology of the immune system. The choice of predictor vari-ables was based on their availability in the HEAD study datasets as noted above. Furthermore, as it became apparent that most (88.7%) patients did not undergo funduscopy, papilloedema was not included as a predictor in the logistic regression because this would have significantly reduced the sample size available for the regression analysis.
For the subgroup without neurological findings, patients with newfocalneurologicalsignsorGCS≤12wereexcluded,aswaspapilloedema. A logistic regression was similarly performed. The regression analysis provided the ORs for serious secondary head-ache of each of the red flag criteria adjusted for other red flags or predictors. Statistical analysis was performed using Stata V.16.1 (College Station, Texas).
Additional statistical analysis The proportions of patients with red flags in the serious and non- serious secondary headache groups were compared using thePearsonχ2orFisher’sexacttestasappropriate.Wealsoperformed an evaluation of the diagnostic accuracy of the red flags using the same method as described by García- Azorín et al.7 Notably, that method excluded patients with missing data as in our study. The sensitivity, specificity, predictive values and area under the receiver operating characteristic curve (AUC) of the SNNOOP10 criteria were reported. The AUC is a measure of how well the criteria discriminate between serious and non- serious secondary headaches.
A sensitivity analysis was performed and explored whether the patient had neurological findings and whether funduscopy was performed. Four groups were analysed: (1) overall cohort including papilloedema (present or absent) as a predictor, (2) overall cohort not including papilloedema, (3) subgroup without neurological findings including papilloedema as a predictor and (4) subgroup without neurological findings not including papilloedema. The analysis regarding neurological findings was planned while the analysis on papilloedema stemmed from the knowledge that most patients did not have a funduscopy.
Sample size No sample size calculation was performed because this was a secondary analysis.
Clinical trial registration The study was registered with the Australian New Zealand Clin-ical Trials Registry (trial number 376695).
Patient and public involvement Patients and the public were not involved in the design or recruit-ment of this study. Results were not disseminated to patients.
RESULTS A total of 5293 patients were included in the HEAD (n=4536) and HEAD- Colombia (n=757) studies. Demographic data of the sample overall are shown in table 1. The breakdown by country is shown in online supplemental table 1. A defined serious head-ache cause was found in 6.1% (323/5293, 95% CI 5.5% to 6.8%; table 2).
Predictors of serious secondary headaches Sample derivation for each of the analyses is shown in figure 2. The presence of red flag criteria in the serious versus non- serious
Table 2 Serious headache causes
HEAD and HEAD- Colombia studies
HEAD study
HEAD- Colombia study
n (%) n (%) n (%)
Neoplasm 58 (18.0) 44 (19.2) 14 (22.2)
Non- subarachnoid haemorrhage intracranial haemorrhage
57 (17.6) 50 (19.2) 7 (11.1)
Meningitis 50 (15.5) 46 (16.9) 6 (9.5)
Subarachnoid haemorrhage 44 (13.6) 34 (13.1) 10 (15.9)
Stroke 44 (13.6) 34 (13.1) 10 (15.9)
Idiopathic intracranial hypertensions 39 (12.1) 26 (10.0) 13 (20.6)
Temporal arteritis 12 (3.7) 12 (4.6) 0
Hydrocephalus 4 (1.2) 4 (1.5) 0
Encephalitis 3 (0.9) 3 (1.2) 0
Vascular dissection 3 (0.9) 2 (0.8) 1 (1.6)
Ventriculoperitoneal shunt complications
3 (0.9) 3 (1.2) 0
Cerebral abscess 2 (0.6) 0 2 (3.2)
Hypertensive crisis 2 (0.6) 2 (0.8) 0
Pregnancy hypertension 2 (0.6) 2 (0.8) 0
Total 323 260 63
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
secondary headache groups is shown in table 3. Key findings were that neurological deficit (new focal neurological signs or GCS≤12;adjustedOR(aOR)6.63,95history of neoplasm (aOR 7.82, 95% CI 4.89 to 12.52) were strongly associated with serious secondary headache diagnosis, with older age (>50 years; aOR 2.00, 95% CI 1.39 to 2.86)
and head trauma (between 3 and 7 days previously; aOR 2.67, 95% CI 1.08 to 6.55) also being significantly associated but less strongly. Notably, sudden onset of headache was not after adjusting for other predictors (aOR 1.43, 95% CI 0.91 to 2.24) and nor was fever (aOR 2.03, 95% CI 0.73 to 5.62). Onset during exertion or sexual activity, pregnancy or puerperium and
Figure2Diagramshowingthepopulationderivationforanalyses.αMaybemorethanonereasonforexclusion.∗ThecalculationofadjustedORsfrom the logistic regression model requires all predictor variables (red flags) to have non- missing data. Papilloedema was excluded as a predictor because data on papilloedema were missing in most (88.7%) patients as funduscopy was not performed. †Cases with missing data to determine whether SNNOOP10 criteria were present (one or more red flags were present) or absent (all red flags were absent). ¶When papilloedema was excluded as a predictor in the analysis, there were less cases with missing data, and so the sample size analysed became larger.
Table 3 Red flags by serious secondary headache diagnosis
Non- serious secondary headaches n=4970
Serious secondary headaches n=323 Adjusted OR* (95% CI)
Fever (T>38°C) 115/4649 2.5% 17/295 5.8% 2.03 (0.73 to 5.62)
Neoplasm 119/2915 4.1% 43/206 20.9% 7.82 (4.89 to 12.52)
Neurological deficit
New focal neurological signs 126/4970 2.5% 50/323 15.5% –
GCS≤126/4451 0.13% 13/303 4.3% –
Any of the above 129/4463 2.9% 58/305 19.0% 6.63 (4.00 to 10.99)
Sudden- onset headache 639/3972 16.1% 62/253 24.5% 1.43 (0.91 to 2.24)
Age >50 years 1480/4970 29.8% 170/323 52.6% 2.00 (1.39 to 2.86)
Precipitated by
Exertion 283/4970 5.7% 18/323 5.6% –
Sexual activity 61/4970 1.2% 4/323 1.2% –
Any of the above 322/4970 6.5% 21/323 6.5% 1.31 (0.64 to 2.68)
Papilloedema 9/539 1.7% 12/61 19.7% –†
Pregnancy or puerperium‡
Pregnancy 114/3308 3.5% 5/179 2.8% –
Puerperium 5/3309 0.15% 1/179 0.56% –
Any of the above 119/3309 3.6% 6/179 3.4% 1.23 (0.29 to 5.24)
Head trauma§ 128/4970 2.6% 10/323 3.1% 2.67 (1.08 to 6.55)
Pathology of immune system¶ 16/4970 0.32% 1/323 0.31% 0.85 (0.10 to 7.17)
*From a multivariate logistic regression analysis, n=2291. †Papilloedema was omitted in the logistic regression model because 88.7% of patients did not have a funduscopy performed. ‡Percentages in females without and with serious secondary headaches, respectively. §Head trauma >2 days and <1 week. ¶Included immunosuppressive drugs (excluding steroids), chemotherapeutic agents, HIV, intravenous drug user and systemic lupus erythematosus.
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
immune suppression were not associated with a serious head-ache diagnosis.
The subgroup analysis for patients who did not have neurolog-ical findings is shown in table 4. In this subgroup, fever (>38°C; aOR 3.27, 95% CI 1.20 to 8.92) was associated with a serious headache diagnosis while other predictors (history of neoplasm, older age and head trauma) also remained significantly associ-ated after adjusting for other predictors.
Diagnostic accuracy in overall cohort including papilloedema as a predictor In the overall cohort, data to determine whether red flag is present (one or more red flags are present) or absent (all red flags are absent) were only available in half the patients (52.7%, 2791/5293). Diagnostic accuracy was calculated with the avail-able data including funduscopy data. The sensitivity of the 10
red flag criteria studied was 96.5% (95% CI 93.4% to 98.4%) but specificity was very low, 5.1% (95% CI 4.3% to 6.0%). AUC was 0.51 (0.50–0.52) . Positive predictive value (PPV) was 9.3% (95% CI 8.2% to 10.4%) and negative predictive value (NPV) was 93.5% (95% CI 88.0% to 97.0%) (table 5).
Diagnostic accuracy in the overall cohort excluding papilloedema as a predictor Funduscopy was not performed in a large proportion of patients (88.7%) resulting in missing data on papilloedema (present or absent). When papilloedema was excluded as a predictor, data to determine whether red flag is present or absent were increased to 71.7% of patients (3781/5293). Diagnostic accu-racy was calculated with the available data. The sensitivity of the remaining nine red flags was 87.5% (95% CI 82.9% to 91.2%) with improved specificity of 31.6% (95% CI 30.1% to 33.2%).
Table 5 Predictive performance of 10 red flag criteria and outcome
SNNOOP10 criteria
Overall cohort Overall cohort—funduscopy excluded
No neurological features subgroup (funduscopy included)
No neurological features subgroup (funduscopy excluded)
Serious headache cause
Non- serious cause Total
Serious headache cause
Non- serious cause Total
Serious headache cause
Non- serious cause Total
Serious headache cause
Non- serious cause Total
Present* 246 2407 2653 237 2400 2637 153 1965 2118 144 1958 2102
Absent† 9 129 138 34 1110 1144 9 123 132 33 1062 1095
255 2536 2791 271 3510 3781 162 2080 2250 177 3020 3197
Sensitivity 96.5% (93.4–98.4%) 87.5% (82.9–91.2%) 94.4% (89.7–97.4%) 81.4% (74.8–86.8%)
Specificity 5.1% (4.3–6.0%) 31.6% (30.1–33.2%) 5.9% (4.9–7.0%) 35.2% (33.5–36.9%)
PPV, % (95% CI) 9.3% (8.2% to 10.4%) 9.0% (7.9% to 10.1%) 7.2% (6.2% to 8.4%) 6.9% (5.8% to 8.0%)
NPV, % (95% CI) 93.5% (88.0% to 97.0%) 97.0% (95.9% to 97.9%) 93.2% (87.5% to 96.8%) 97.0% (95.78% to 97.9%)
*At least one red flag. †All were recorded as absent, that is, not as missing. NPV, negative predictive value; PPV, positive predictive value.
Table 4 Subgroup analysis for patients without focal neurological signs, confusion or reduced level of consciousness (GCS<15)
Non- serious secondary headaches n=4269
Serious secondary headachesn=220 Adjusted OR* (95% CI)
Fever (T>38°C) 99/4126 2.4% 13/211 6.2% 3.27 (1.20 to 8.92)
Neoplasm 111/2576 4.3% 30/143 21.0% 7.27 (4.34 to 12.19)
Sudden- onset headache 485/3477 14.0% 33/177 18.6% 1.07 (0.61 to 1.88)
Age >50 years 1227/4269 28.7% 100/220 45.5% 1.88 (1.25 to 2.82)
Precipitated by
Exertion 256/4269 6.0% 17/220 7.7% –
Sexual activity 53/4269 1.2% 3/220 1.4% –
Any of the above 289/4269 6.8% 19/220 6% 1.32 (0.62 to 2.80)
Papilloedema 8/471 1.7% 12/46 26.1% –†
Pregnancy or puerperium‡
Pregnancy 98/2875 3.4% 4/134 3.0% –
Puerperium 4/2876 0.14% 1/134 0.75% –
Any of the above 102/2876 3.6% 5/134 3.7% 1.41 (0.33 to 6.01)
Head trauma§ 113/4269 2.7% 8/220 3.6% 2.66 (1.01 to 6.97)
Pathology of immune system¶ 16/4269 0.37% 0/220 0% –**
*From a multivariate logistic regression analysis, n=2137. †Papilloedema was omitted in the logistic regression model because 88.7% of patients did not have a funduscopy performed. ‡Percentages in females without and with serious secondary headaches, respectively. §Head trauma >2 days and <1 week. ¶Included immunosuppressive drugs (excluding steroids), chemotherapeutic agents, HIV, intravenous drug user and systemic lupus erythematosus. **Omitted because of no pathology of immune system in serious secondary headache group.
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
AUC was 0.60 (95% CI 0.57 to 0.62) (online supplemental figure 1). PPV was 9.0% (95% CI 7.9% to 10.1%) and NPV was 97.0% (95% CI 95.9% to 97.9%) (table 5).
Diagnoses that were missed by the red flags were IIH (15), neoplasm (6), viral meningitis (6), SAH (2), stroke (2), ICH (not SAH) (1), VP shunt complication (1) and hydrocephalus (1), constituting a total 34 (12.5%) of serious diagnosis cases.
Diagnostic accuracy in subgroup without neurological findings including papilloedema as a predictor In this subgroup without neurological findings, data to determine whether red flag is present or absent were again only available in half the patients (50.1%, 2250/4489). Diagnostic accuracy was calculated with the available data including funduscopy data. The sensitivity of the 10 red flag criteria was 94.4% (95% CI 89.7% to 97.4%) and specificity was 5.9% (95% CI 4.9% to 7.0%). AUC was 0.52 (95% CI 0.48 to 0.52). PPV was 7.2% (95% CI 6.2% to 8.4%) and NPV was 93.2% (95% CI 87.5% to 96.8%) (table 5).
Diagnostic accuracy in subgroup without neurological findings excluding papilloedema as a predictor When papilloedema was excluded, data to determine whether red flag is present or absent were available in 71.2% (3197/4489). Diagnostic accuracy was calculated with the available data. The sensitivity of the remaining nine red flags was 81.4% (95% CI 74.8% to 86.8%) and specificity was 35.2% (95% CI 33.5% to 36.9%). AUC was 0.58 (95% CI 0.55 to 0.61) (online supple-mental figure 2). PPV was 6.9% (95% CI 5.8% to 8.0%) and NPV was 97.0% (95% CI 95.8% to 97.9%) (table 5).
Diagnoses that were missed by the red flags (excluding fundus-copy) were similar to the group described above with the excep-tion of a case of ICH.
DISCUSSION Summary This analysis of a large multinational study addressed the associ-ation between commonly accepted red flags for a serious head-ache diagnosis in the specific setting of ED. Its findings challenge the utility of some of the SNNOOP10 criteria in that setting.
For the overall cohort, new focal neurological signs and history of neoplasm were most strongly associated with serious secondary headache diagnosis, with older age and head trauma also being significantly associated but less strongly. Papilloedema was also associated with serious causes in the univariate anal-ysis, but small numbers of patients with funduscopy precluded its inclusion in the multivariate analysis. Interestingly, headache of sudden onset was not associated with serious causes after adjusting for other predictors.
Perhaps of most relevance to clinical practice in the ED is the subgroup analysis of patients who presented with normal conscious state and without new neurological features (other than headache) or new confusion as these patients have higher diagnostic uncertainty. In that analysis, history of neoplasm, age >50 years and head trauma were again associated with a serious headache diagnosis along with fever. In particular, sudden (thun-derclap) headache was not, with and without adjusting for other predictors.
Our findings highlight a key problem with the validation of the red flag approach to identification of serious headache in the ED setting. In particular, the broad range of headache causes and the association of some red flags with specific rare conditions are likely to result in poor predictive performance of some red flags
in large and diverse headache populations. That said, the high sensitivity of the red flags as a group may support the use of red flags in conjunction with clinical gestalt. The relative diagnostic accuracy of clinical gestalt, a red flag approach or a combination of these has not yet been investigated.
The low rate of funduscopy is a challenge to the validity of our results. It, however, reflects the reality of contemporary emergency medicine practice.1 With ready access to advanced imaging in most developed countries, the additional benefit of funduscopy can be questioned. Importantly, there is evidence that even when funduscopy is performed in ED, it has poor accu-racy for detection of serious conditions—as low as 0% in one study.8–10 This emphasises the importance of identifying objec-tive, reliable and easily assessable criteria other than funduscopy that accurately predict a serious headache cause.
Comparison to previous literature In this study, sensitivity of the combined red flag criteria was lower than previously reported and specificity of the criteria was also low. One previous study of the red flag criteria has shown associations between immunosuppression and older age with secondary headache aetiologies but did not confirm a similar association for sudden onset of headache or abnormal neurolog-ical examination.6 Another study reported that all patients with study- defined high- risk headaches had at least one SNNOOP10 criterion.7 In that study, the criteria significantly associated with high- risk headache were older age, post- traumatic onset, neuro-logical deficit or dysfunction, and neoplasm in history.7 That study also noted that most of the criteria had low specificity for high- risk headache.7 As noted above, the number of patients in both of these studies was much smaller than our cohort.
Regarding sensitivity of the red flag criteria, a study of the SNOOP4 reported a sensitivity of 77.8% with specificity of 73%.11 Only one small study (of 100 patients) found sensitivity of 100% but had significant selection bias due to inclusion of urgent triage categories only which may have overestimated the sensitivity.7 In the experience of our research group, patients who are neurologically normal with normal vital signs are more likely to be assigned low triage categories so would have been excluded from that analysis. Unpublished data from this analysis found that approximately 14% of serious headache occurred in patients with lower triage categories.
Onset during exertion including sexual activity, pregnancy and immune suppression were not associated with a serious head-ache diagnosis. A subgroup analysis of the pregnant subgroup in the HEAD study has previously been published to support this finding.12
It should be noted that the previous research studies were both single- centre studies with much smaller sample sizes than this study.6 7 They collected data on patient and headache features prospectively and using structured tools. All patients were then assessed by a neurologist with ready access to advanced neuro-imaging. It is not clear if study neurologists were blinded to the study hypotheses. This is different from the real world of most EDs where assessments are mainly performed by ED doctors of varying seniority and experience and without ready access to all neuroimaging modalities.
Strengths and weaknesses The strengths of this study are that it represents a real- world ED cohort of patients presenting with headache, has large numbers and was carried out in ED in several countries with different healthcare models.
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
Limitations include that classification of headache as the main symptom and ED diagnosis were based on clinician judge-ment. This has been shown to be difficult to classify accurately in ED.13 Although patients were identified prospectively, some data were collected retrospectively with the inherent risks that impose, including of missing data.13 With the exception of some Queensland sites and the UK where some form of consent was required, participating institutions were instructed to include all patients presenting with headache within the enrolment period, but some patients may have been missed. Resource did not allow verification of this. That said, given the high number of partic-ipating patients, it is unlikely that missed patients at individual EDs would have introduced systematic bias. The design of the study and resource limitations precluded assessment of inter- rater reliability of data collection. Diagnosis was as determined by the ED physician at the end of the ED phase of care. It is possible that some patients may have had further investigations after the ED phase of care which may have identified an alter-native diagnosis. Similarly, the nature of ED practice precludes validation of diagnoses. The hospitals were mostly located in developed countries so findings may not be generalisable to the developing world.
Regarding the pooling of data, the parent study was a 1- month snapshot while that HEAD- Colombia study included data collected over a longer period. The HEAD- Colombia study site is a specialist neurological referral centre. This may have resulted in different ED attendance patterns for patients with headache. Not all patients had complete data. We chose to exclude patients with missing data. We chose not to use other approaches such as imputation. We considered the risk of bias from using complete data cases only was less than that of using other statistical approaches to missing data.
Implications for clinical practice and research The high sensitivity of the SNNOOP10 criteria suggests that they are useful to identify patients at higher risk of a serious secondary headache cause. Sensitivity, however, was not high enough that their absence alone should not be used to determine whether further investigation is required. The low rate of funduscopy and its reported inaccuracy suggest that new and more accurate ways of examining the optic fundus may be needed. Further research is needed, especially focusing on patients with no new neuro-logical features, to identify clinical features predictive of serious secondary headache diagnoses.
CONCLUSION In this ED- based study, sensitivity of the red flag criteria was lower than previously reported and specificity of the criteria was also low, more so for patients with no new neurological features (other than headache) and in patients in whom funduscopy was not performed. Further research is needed, especially focusing on patients with no new neurological features, to identify clin-ical features predictive of serious secondary headache diagnoses.
Author affiliations 1Emergency and Trauma Centre, Royal Brisbane and Women’s Hospital, Brisbane, Queensland, Australia 2Joseph Epstein Centre for Emergency Medicine Research, Western Health, Footscray, Victoria, Australia 3Department of Critical Care, The University of Melbourne, Melbourne, Victoria, Australia 4Emergency Medicine Department, National University Hospital, Singapore 5Emergency and Children’s Services, The Prince Charles Hospital, Brisbane, Queensland, Australia 6Department of Medicine, University of Queensland, Brisbane, Queensland, Australia
7Department of Emergency Medicine, Gold Coast University Hospital, Southport, Queensland, Australia 8Emergency Department, Salford Royal NHS Foundation Trust, Salford, UK 9Emergency Medicine Department, Tours University Hospital, Tours, France 10Instituto Neurológico de Colombia, Medellin, Colombia 11Department of Emergency Medicine, Gazi University Faculty of Medicine, Ankara, Turkey 12Department of Neurology, Western Health, Footscray, Victoria, Australia 13School of Medicine, University of Auckland, Auckland, New Zealand 14Auckland City Hospital, Auckland District Health Board, Auckland, New Zealand 15Emergency Medicine, Chinese University of Hong Kong - Prince of Wales Hospital, Sha Tin, Hong Kong 16Emergency Department, Manchester University NHS Foundation Trust, Manchester, UK 17Division of Cardiovascular Sciences, University of Manchester, Manchester, UK 18Trainee Emergency Research Network, London, UK 19North Bristol NHS Trust, Bristol, UK
X Anne- Maree Kelly @kellyam_jec, Daniel Horner @rcemprof, Richard Body @ richardbody and Tom Roberts @DrTomRoberts
Acknowledgements HEAD Study Steering Group: Anne Maree Kelly; Kevin H Chu; Win Sen Kuan; Gerben Keijzers; Frances B Kinnear; Mehmet A Karamercan; Sharon Klim; Tissa Wijeratne; Sinan Kamona; Colin A Graham; Richard Body; Tom Roberts; Daniel Horner; Said Laribi. HEAD- Colombia: Alejandro Cardozo- Ocampo; Valentina Jaramillo; Paola Parra. HEAD Study Group: Catherine Lunter (Coffs Harbour Hospital, New South Wales, Australia); Rochelle Facer (Concord Repatriation Hospital, New South Wales, Australia); David Thomson (Port Macquarie Base and Kempsey District Hospitals, New South Wales, Australia); Robert Day (Royal North Shore Hospital, New South Wales, Australia); Greg McDonald (Sydney Adventist Hospital, New South Wales, Australia); Sarah Jones (Tamworth Regional Hospital, New South Wales, Australia); Julian Cochrane (Orange Base Hospital, New South Wales, Australia); Stephen Gourley (Alice Springs Hospital, Northern Territory, Australia); Mark Ross and Vinay Gangathimmaiah (Royal Darwin Hospital, Northern Territory, Australia); Kim Hansen (St Andrew’s War Memorial Hospital, Queensland, Australia); Frances B Kinnear (The Prince Charles Hospital, Queensland, Australia); Gerben Keijzers (Gold Coast University Hospital, Queensland, Australia); Kevin Chu (Royal Brisbane and Women’s Hospital, Queensland, Australia); Paul Bowe (Robina Hospital, Queensland, Australia); Raymund de la Cruz (Lyell McEwin and Modbury Hospitals, South Australia, Australia); Daniel Haustead (The Queen Elizabeth and Royal Adelaide Hospitals, South Australia, Australia); Jean Moller (University Hospital Geelong, Victoria, Australia); Katie Walker (Cabrini Malvern, Victoria, Australia); Richard D Smith (Bendigo Health, Victoria, Australia); Ron Sultana (Epworth Healthcare, Victoria, Australia); John Pasco (Werribee Mercy Hospital, Victoria, Australia); Neil Goldie and Andis Graudins (Monash Health, Victoria, Australia); Rosamond Dwyer (Peninsula Health, Victoria, Australia); George Plunkett (Melbourne Health, Victoria, Australia); Anne- Maree Kelly (Western Health, Victoria, Australia); Hugh Mitenko (WA Country Health Service, Western Australia); Michael Lovegrove (Joondalup Health Campus, Western Australia); Ben Smedley (Rockingham General Hospital, Western Australia); Colin A Graham and Ling Yan Leung (Prince of Wales Hospital, Hong Kong SAR), Win Sen Kuan and Ying Wei Yau (National University Hospital, Singapore); Wei Ming Ng (Ng Teng Fong General Hospital, Singapore); Ranjeev Kumar (Khoo Teck Puat Hospital, Singapore); Dennis Wen Jie Chia (Sengkang General Hospital, Singapore); Said Laribi (CHU Tours, Tours, France); Mounir Hilal and Rarthtana Mil (CH Vendôme, France); Audrey Gerineau (CHR Orléans, France); Matthew J Reed (Emergency Medicine Research Group Edinburgh (EMERGE), Royal Infirmary of Edinburgh, UK); Daniel Horner (Salford Royal NHS Foundation Trust, Salford, UK); Edward Carlton and Tom Roberts (North Bristol NHS Trust, UK); Girish Boggaram and Jayne Foot (Musgrove Park Hospital, Taunton, UK); Andy Appleboam, Rachel Goss and Hamza Malik (Royal Devon and Exeter NHS Foundation, UK); Richard Body (Manchester Royal Infirmary, Manchester, UK); John- Paul Williamson (Royal Oldham Hospital, Oldham, UK); Adela Golea and Sonia Luka (University County Hospital Cluj- Napoca, Romania); Huseyin Avni Demir (University of Health Sciences Mehmet Akif Inan Training and Research Hospital, Department of Emergency Medicine, Şanlıurfa, Turkey); Şafak Öner Gülpinar (Tokat Erbaa Government Hospital, Tokat, Turkey); Lale Tolu (Bursa Çekirge Government Hospital of Emergency Service, Bursa, Turkey); Muhammet Hacimustafaoğlu (Hakkari Yuksekova Government Hospital, Hakkari, Turkey); Mehmet A Karamercan (Gazi University Faculty of Medicine, Department of Emergency Medicine, Ankara, Turkey); Elif Çelikel (Numune Research and Training Hospital, Department of Emergency Medicine, Ankara, Turkey); Çilem Çaltili (University of Health Sciences Bağcılar Training and Research Hospital, Department of Emergency Medicine, Istanbul, Turkey); Selahattin Gürü (Yıldırım Beyazıt University Faculty of Medicine, Department of Emergency Medicine, Ankara, Turkey); Gülşah Yavuz (Antalya Ataturk Government Hospital of Medicine, Department of Emergency Medicine, Antalya, Turkey); Franck Verschuren (Institute of Experimental and Clinical Research, Emergency Department, Saint- Luc University Hospital, Brussels, Belgium); Christopher Ramos (Emergency Department, Saint- Luc University Hospital, Brussels, Belgium); Paule Denoel and Nicolas Wilmet (Saint Michel, Clinique de l’Europe,
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
Etterbeek, Brussels); Michael Vandoorslaert and Alessandro Manara (Saint Elisabeth, Clinique de l’Europe, Uccle, Brussels); Adeline Higuet (CHR Hal, Belgium); Amichai Sheffy (Tel- Aviv Sourasky Medical Center, Israel); Sinan Kamona and Peter Jones (University of Auckland, School of Medicine, Auckland, New Zealand); Mai Nguyen (Wellington Hospital, Wellington, New Zealand); Anne Clarke (Hutt Valley Hospital, Lower Hutt, New Zealand); Sierra Beck (Dunedin Hospital, Dunedin, New Zealand); Andrew Munro (Nelson Hospital, Nelson, New Zealand); Kim M Yates (North Shore and Waitakere Hospitals, Waitematā District Health Board, New Zealand); James Weaver (Christchurch Hospital, Christchurch, New Zealand); Deborah Moore and Stuart Innes (Tauranga Hospital, Tauranga, New Zealand); Karina Walters (Taranaki District Health Board, New Zealand). Coordinating Centre, Western Health, Victoria, Australia: Anne- Maree Kelly, Sharon Klim and Kerrie Russell.
Contributors KC and A- MK had the concept for the study. The authors codesigned the study and facilitated data collection. KC undertook the analysis and is the guarantor for this work. All authors had input into interpretation of the results. A- MK drafted the manuscript. All authors contributed to refinement of the manuscript.
Funding The Royal College of Emergency Medicine (UK) provided part funding (G/2018/1).
Competing interests RB is deputy editor of EMJ.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Patient consent for publication Not applicable.
Ethics approval The lead ethics approvals were from Melbourne Health Human Research Ethics Committee (HREC/43148/MH- 2018) and Instituto Neurologico de Colombia–Medelli n (PR0107). Approval was obtained for other participating sites according to local requirements.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. Data may be available subject to HREC approval.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer- reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
ORCID iDs Anne- Maree Kelly http://orcid.org/0000-0002-4655-5023 Daniel Horner http://orcid.org/0000-0002-0400-2017 Richard Body http://orcid.org/0000-0001-9089-8130 Tom Roberts http://orcid.org/0000-0003-4991-974X
REFERENCES 1 Kelly AM, Kuan WS, Chu KH, et al. Epidemiology, investigation, management, and
outcome of headache in emergency departments (HEAD study)- a multinational observational study. Headache 2021;61:1539–52.
2 Cardozo A, Jaramillo V, Parra P, et al. Epidemiology of headache in a neurological emergency department in medellin, Colombia. Headache Med 2023;14:43–8.
3 National Clinical Guideline Centre (UK). Headaches: diagnosis and management of headaches in young people and adults. NICE clinical guidelines, no. 150. London Royal College of Physicians (UK); 2012.
4 American Headache Society. Red flags in headache—what if it isn’t migraine? Available: https://americanheadachesociety.org/news/red-flags-in-headache-what-if- it-isnt-migraine/ [Accessed 4 Jan 2023].
5 Do TP, Remmers A, Schytz HW, et al. Red and orange flags for secondary headaches in clinical practice: SNNOOP10 list. Neurology 2019;92:134–44.
6 Munoz- Ceron J, Marin- Careaga V, Peña L, et al. Headache at the emergency room: etiologies, diagnostic usefulness of the ICHD 3 criteria, red and green flags. PLoS One 2019;14:e0208728.
7 García- Azorín D, Abelaira- Freire J, González- García N, et al. Sensitivity of the SNNOOP10 list in the high- risk secondary headache detection. Cephalalgia 2022;42:1521–31.
8 Biousse V, Bruce BB, Newman NJ. Ophthalmoscopy in the 21st century: the 2017 H. Houston merritt lecture. Neurology 2018;90:167–75.
9 Petrushkin H, Barsam A, Mavrakakis M, et al. Optic disc assessment in the emergency department: a comparative study between the panoptic and direct ophthalmoscopes. Emerg Med J 2012;29:1007–8.
10 Dunn HP, Browning SD, Thomson D, et al. Impact on patient management of non- mydriatic fundus photography compared to direct ophthalmoscopy in a regional Australian emergency department. Emerg Med Australas 2022;34:186–93.
11 Wongtanasarasin W, Wittayachamnankul B. Clinical availability of SNOOP4 in acute non- traumatic headache patients admitted to the emergency department. Hong Kong J Emerg Med 2022;29:161–7.
12 Kelly AM, Chu KH, Kuan WS, et al. Is headache during pregnancy a higher risk for serious secondary headache cause? A HEAD study report. Emerg Med Australas 2022;34:629–31.
13 Gilbert EH, Lowenstein SR, Koziol- McLain J, et al. Chart reviews in emergency medicine research: where are the methods. Ann Emerg Med 1996;27:305–8.
by copyright. on A
pril 30, 2024 at B en G
urion U ni M
A LM
A D
C onsortia. P
rotected http://em
j.bm j.com
/
erg M
ed J: first published as 10.1136/em erm
ed-2023-213461 on 24 A pril 2024. D
ow nloaded from
03/17/2020 1:10pm projectredcap.org
Confidential Headache in Emergency Departments (Head Study)
Page 1
Case Number and Demographics
Record ID __________________________________
Select your Country 1 Australia 2 New Zealand 3 Hong Kong 4 Singapore 5 France 6 United Kingdom 7 Israel 8 Belgium 9 Turkey 10 Romania 11 Ireland 12 Switzerland
Australian State 1 ACT 2 NSW 3 NT 4 QLD 5 SA 6 TAS 7 VIC 8 WA
NSW site 1 Blacktown 2 Calvary Mater Newcastle 3 Canterbury 4 Coffs Harbour 5 Concord Repatriation General 6 Kempsey District 7 Lismore Base 8 Mt Druitt 9 Orange Base 10 Port Macquarie 11 Royal North Shore 12 Shoalhaven 13 Sydney Adventist 14 Tamworth 15 The Maitland
(Select from drop down list)
ACT site 1 Calvary Public Bruce
NT Site 1 Alice Springs 2 Royal Darwin
QLD site 1 Cairns 2 Gold Coast 3 Mater Adult Public 4 Mt Isa 5 Queen Elizabeth II Jubilee 6 Robina 7 St Andrew's War Memorial 8 Royal Brisbane and Women's 9 The Prince Charles
03/17/2020 1:10pm projectredcap.org
Confidential Page 2
SA Site 1 Calvary Wakefield 2 Flinders Medical Centre 3 Lyell McEwin 4 Modbury Public 5 Royal Adelaide 6 The Queen Elizabeth
TAS Site 1 North West Regional (Burnie) 2 Royal Hobart
VIC site 1 Austin Health 2 Bendigo 3 Cabrini (Malvern) 4 Casey (Monash Health) 5 Clayton (Monash Health) 6 Dandenong (Monash Health) 7 Epworth Richmond 8 Footscray (Western Health) 9 Frankston (Peninsula Health) 10 Royal Melbourne 11 St John of God (Geelong) 12 Sunshine (Western Health) 13 University Hospital Geelong (Barwon) 14 Mercy Health
WA site 1 Bunbury Regional 2 Joondalup Health 3 Sir Charles Gairdner 4 St John of God (Midland) Public 5 Rockingham General
Hong Kong Site 1 Prince of Wales
Singapore Site 1 Khoo Teck Puat 2 National University 3 Ng Teng Fong General 4 Sengkang General
France site 1 CHU Tours 2 CH Le Mans 3 CH Vendome 4 CH Chinon 5 CHR Orleans
Belgium Site 1 UC Louvain Brussels Belgium University 2 Cliniques Universitaires Saint-Luc 3 Cliniques de l'Europe- sainte-Elisabeth 4 Cliniques de l'Europe- St-Michel 5 CHU de Charleroi 6 CHU Liège 7 CHR Hal 8 Cliniques Saint-Jean
Ireland Site 1 St Vincents Dublin
03/17/2020 1:10pm projectredcap.org
Confidential Page 3
United Kingdom Site 1 Royal Infirmary of Edinburgh 2 Salford Royal NHS Foundation Trust 3 North Bristol NHS Trust 4 Taunton and Somerset NHS Foundation Trust ( Musgrove Park site) 5 Royal Devon and Exeter NHS Foundation Trust 6 Manchester Royal Infirmary 7 Cardiff and Vale University Health Board (UHB) 8 Royal Oldham
Romania Site 1 County Hospital Cluj Cluj Napoca
Turkey Site 1 Gazi University School of Medicine 2 Ankara Numune Education and Research Hospital 3 Istanbul Bagcilar Education and Research Hospital 4 Ankara Yildirim Beyazit Faculty of Medicine (University) 5 Sanliurfa Mehmet Akif Inan Education and Research Hospital 6 Tokat Erbaa Government Hospital 7 Bursa Cekirge Government Hospital 8 Hakkari Yuksekova Government Hospital 9 Antalya Ataturk Government Hospital
Israel Site 1 Tel-aviv Sourasky Medical Center
New Zealand Site 1 Auckland City 2 North Shore 3 Waitakere 4 Tauranga 5 Wellington Regional 6 Christchurch 7 Dunedin 8 Hutt Valley 9 Middlemore 10 Nelson 11 Rotorua 12 Waikato 13 Taranaki Base
03/17/2020 1:10pm projectredcap.org
Confidential Page 4
Ethnicity (NZ only) 1 NZ European 2 Australian 3 European NFD 4 NZ Maori 5 Samoan 6 Tongan 7 Cook Island Maori 8 Pacific Islander
NFD 9 African 10 American 11 Asian NFD 12 Chinese 13 Fijian 14 Fijian Indian 15 Indian 16 Latin American/Hispanic 17 Middle Eastern 18 Niuean 19 Southeast Asian 20 Tokelauan 21 Other 22 Unknown
(Check all boxes that apply (for NZ sites only; required under NZ national ethics approval guidelines))
Age __________________________________
Gender 1 Male 2 Female 3 Transgender 4 Unknown
Known Current Pregnancy 1 No 2 Yes
Referred by 1 Self 2 GP/other doctor (if not documented assume self)
Mode of Arrival 1 Private Transport/Self 2 Ambulance 3 Other
Triage Category 1 Immediate 2 Urgent (2 and 3 on a five point scale) 3 Non Urgent (4 and 5 on a five point scale)
03/17/2020 1:10pm projectredcap.org
Confidential Headache in Emergency Departments (Head Study)
Page 5
Past Medical History and Regular Medication
Known Past Medical History 1 No (if not documented assume No) 2 Yes
(If not documented select NO)
No Yes History of recurrent headache (migraine excluded)
Previous migraine diagnosis Previous cluster headache diagnosis
Previous tension headache diagnosis
Previous stroke/ TIA Serious intracranial injury - EDH, SDH, traumatic SAH, cerebral contusion requiring hospital admission/ neurosurgery
Presence of a ventriculo-peritoneal shunt
Malignant Intracranial neoplasm - primary
Malignant Intracranial neoplasm - secondary
Intracranial neoplasm - unknown benign v malignant
Known benign intracerebral tumour e.g. Meningioma
Non-cerebral malignancy without known intracranial secondary neoplasm
Subarachnoid haemorrhage Intracranial aneurysm without SAH
Intracranial hypertension Known Intracranial vascular abnormality e.g.AVM
Other Past Medical History (not listed above and you consider relevant to the cause of headache)
Other Past Medical History __________________________________
03/17/2020 1:10pm projectredcap.org
Confidential Page 6
Regular Medications Taken 1 No 2 Yes
Regular Medications (If information is NOT documented select NO) No Yes
Triptan Beta-blockers - propranolol, metoprolol, atenolol, bisoprolol, timolol, etc
Pizotifen (Sandomigran) Topiramate (Topamax) Tricyclic antidepressants -amitriptyline, nortriptyline, etc
Sodium valproate Candesartan Verapamil Botulinum toxin Anticoagulants - Novel Oral Anticoagulants (NOAC), warfarin, Vit K antagonist
Long term use of codeine preparations
Other opioids
03/17/2020 1:10pm projectredcap.org
Confidential Headache in Emergency Departments (Head Study)
Page 7
Clinical History and Clinical Examination
Duration of Symptoms 1= < 24 hours 2= 1-3 days 3= >3 days 4= Unknown
Onset of Symptoms 1 Gradual 2 Sudden/Thunderclap (peaking instantly or almost) 3 Peak within 1 hour but not instant 4 Unknown
Location of Headache 1 Generalized 2 Unilateral 3 Unclear
Severity 1 Mild (pain score up to 3/10) 2 Moderate (pain score 4-7/10) 3 Severe (pain score 8 or more/10) 4 Unclear
Worst headache ever? 1 No 2 Yes
(If not documented select NO)
Head Trauma within the last week 1 No 2 Yes
Relationship to exertion 1 No 2 Yes
(If not documented select NO)
Relationship to sexual activity 1 No 2 Yes
(If not documented select NO)
Reported neck pain or stiffness 1 No 2 Yes
(If not documented select NO)
Nausea or vomiting 1 No 2 Yes
(If not documented select NO)
Syncope/ loss of consciousness 1 No 2 Yes
(If not documented select NO)
Photophobia. Reported by patient. 1 No 2 Yes
(If not documented select NO)
New limb weakness transient or current. Reported by 1 No patient. 2 Yes
(If not documented select NO)
03/17/2020 1:10pm projectredcap.org
Confidential Page 8
New limb paraesthesia transient or current. Reported 1 No by patient. 2 Yes
(If not documented select NO)
New speech difficulty - including slurred speech, 1 No inability to speak, etc. Reported by patient. 2 Yes
(If not documented select NO)
New reported visual disturbance - transient or 1 No ongoing. Reported by patient. 2 Yes
(If not documented select NO)
Subjective fever or rigors. Reported by patient. 1 No 2 Yes
(If not documented select NO)
Rash. Reported by patient. 1 No 2 Yes
(If not documented select NO)
Current or recent Intravenous drug use 1 No 2 Yes
(If not documented select NO)
Medication Taken Pre- ED (this episode) - must 1 No specify to have been self administered by patient 2 Yes
No Yes Paracetamol (pre-ED self administered)
Aspirin (pre-ED self administered)NSAID, excluding Aspirin (pre-ED self administered)
Codeine containing preparation (pre-ED self administered)
Triptan (pre-ED self administered)Oxycodone (e.g. endone, oxycontin, oxynorm, targin) (pre-ED self administered)
Tramadol (pre-ED self administered)
Other Opiate (pre-ED self administered)
Antiemetic-metoclopramide, prochlorperazine, ondansetron (pre-ED self administered)
03/17/2020 1:10pm projectredcap.org
Confidential Page 9
Other medication to treat headache (pre-ED self administered)
Pre ED medications to treat headache or cause of headache- Specify other type not previously listed __________________________________
(specify other medication if applicable)
Ambulance Pre Hospital Medication Administered 1 No 2 Yes 3 Not documented
(This refers to medications administered to treat headache or presumed cause of headache. Must specify medication administered by Ambulance Team)
No Yes Paracetamol (in ambulance) Aspirin (in ambulance) NSAID, excluding Aspirin (in ambulance)
Codeine containing preparation (in ambulance)
Triptan (in ambulance) Oxycodone (e.g. endone, oxycontin, oxynorm, targin) (in ambulance)
Tramadol (in ambulance) Fentanyl (in ambulance) Oramorph (in ambulance) Morphine Sulphate IV (in ambulance)
Other Opiate (in ambulance) Antiemetic-metoclopramide, prochlorperazine, ondansetron (in ambulance)
Methoxyflurane (in ambulance) Antibiotics (in ambulance) Other medication to treat headache or presumed cause of headache (in ambulance)
Other Medications given by Ambulance to treat headache or presumed cause of headache. Please __________________________________ specify type (specify other medication if applicable)
Clinical Examination in ED Pulse Rate __________________________________
(FIRST RECORDED IN EMERGENCY DEPARTMENT)
03/17/2020 1:10pm projectredcap.org
Confidential Page 10
Clinical Examination in ED Systolic BP __________________________________
(FIRST RECORDED IN EMERGENCY DEPARTMENT)
Clinical Examination in ED 1 No TEMPERATURE TAKEN 2 Yes (recorded numerically Celcius)
Clinical Examination in ED 1 AFEBRILE / NO FEVER Temperature recorded 2 FEBRILE / FEVER AFEBRILE / NO FEVER 3 UNKNOWN FEBRILE / FEVER (No numerical temperature recorded, but history
does specify temperature in words)
Clinical Examination in ED Temperature (Celsius) __________________________________
(FIRST RECORDED IN EMERGENCY DEPARTMENT)
Clinical Examination in ED 1 Known Is GCS score known 2 Unknown
GCS- Eye __________________________________
GCS Verbal __________________________________
GCS Motor __________________________________
GCS Overall __________________________________
Clinical Examination in ED 1 No Rash (observed by Clinician) 2 Yes
(If not documented select NO)
Clinical Examination in ED 1 No Confusion (observed by Clinician) 2 Yes
(If not documented select NO)
Clinical Examination in ED 1 No Meningism 2 Yes
(If not documented select NO)
Clinical Examination in ED 1 No Limited Neck Flexion (on examination) 2 Yes
(If not documented select NO)
Clinical Examination in ED 1 No New Focal Neurological Signs 2 Yes
(If not documented select NO)
03/17/2020 1:10pm projectredcap.org
Confidential Page 11
New Focal Neurological Sign 1 Isolated speech deficit 2 Isolated unilateral limb weakness 3 Speech deficit and limb weakness 4 Incoordination/cerebellar signs 5 Other
Describe Other New Focal Neurological Sign __________________________________
Clinical Examination in ED 1 No New Vision Defect 2 Yes
(If not documented select NO)
Clinical Examination in ED 1 Not done Ophthalmoscopy Findings 2 Normal
3 Papilloedema 4 Other (specify)
(If not documented select NO)
Ophthalmoscopy Findings (specification of other findings) __________________________________
03/17/2020 1:10pm projectredcap.org
Confidential Headache in Emergency Departments (Head Study)
Page 12
Investigations
White Cell Count Done 1 No 2 Yes
White Cell Count x10-9/L __________________________________
Neutrophil Count Done 1 No 2 Yes
Neutrophil Count (x10-9/L) __________________________________
C-Reactive Protein Done 1 No 2 Yes
C-Reactive Protein unit of measure mg/L micromol/L
(Select the unit of measure for the C- Reactive Protein Value to be inserted below)
C-Reactive Protein __________________________________
Lumbar Puncture Performed 1 No 2 Yes
Lumbar Puncture Results 1 Normal 2 Indicative of infection on microscopy 3 Indicative of SAH (red cell count or xanthochromia) 4 Indicative of raised intracranial pressure 5 Inconclusive
CT Scan Performed 1 No 2 Yes
CT Scan Result 1 Normal 2 Abnormal
CT Abnormality 1 SAH 2 Other bleed 3 Abscess 4 Neoplasm 5 Other (free text describe)
CT Abnormality (OTHER) description __________________________________
MRI Performed 1 No 2 Yes
MRI Result 1 Normal 2 Abnormal
03/17/2020 1:10pm projectredcap.org
Confidential Page 13
MRI Abnormality 1 Bleed 2 Abscess 3 Neoplasm 4 Other (describe below)
MRI Abnormality (OTHER) description __________________________________
CT Angiography Performed 1 No 2 Yes
CT Angiography Result 1 Normal 2 Abnormal
CT Angiography Abnormality 1 Aneurysm with bleed 2 Aneurysm without bleed 3 No aneurysm 4 Other (free text describe)
CT Angiography (Other) description __________________________________
Other Imaging Performed 1 No 2 Yes
Other Imaging (specify what type of imaging and provide results description) __________________________________
03/17/2020 1:10pm projectredcap.org
Confidential Headache in Emergency Departments (Head Study)
Page 14
ED Treatment and Intervention
Medication to treat headache or cause of headache 1 No given in ED 2 Yes
Medications given after the initial clinical 1 No assessment (including nurse-initiated medications) 2 Yes
No Oral Parenteral Paracetamol administered in ED Aspirin administered in ED NSAID (other than Aspirin) administered in ED
Codeine containing compounds administered in ED
Triptan administered in ED Oxycodone administered in ED Pethidine/Meperidine administered in ED
Other Opioid administered in ED Chlorpromazine Infusion administered in ED
Metoclopramide administered in ED
Ondansetron administered in ED Prochlorperazine administered in ED
Droperidol/ Haloperidol administered in ED
Ergot Alkaloids administered in ED
Corticosteroid administered in EDAntibiotic/ Antiviral agent administered in ED
Other Medication administered in ED to treat headache or cause of headache
OTHER ED Medication. Please specify __________________________________
Treatment in ED after initial clinical assessment 1 No 2 Yes
Treatment in ED 1 No Oxygen Therapy 2 Yes
(after initial clinical assessment )
03/17/2020 1:10pm projectredcap.org
Confidential Page 15
Treatment in ED 1 No Acupuncture 2 Yes
(after initial clinical assessment)
Treatment in ED 1 No Intravenous fluids (not part of a drug infusion) 2 Yes
(after initial clinical assessment)
Follow-up Medications given > 30 minutes after 1 No initial medications 2 Yes
03/17/2020 1:10pm projectredcap.org
Confidential Page 16
No Oral Parenteral Paracetamol administered in ED - more than 30 mins after primary treatment
Aspirin administered in ED -more than 30 mins after primary treatment
NSAID (other than Aspirin) administered in ED - more than 30 mins after primary treatment
Codeine containing compounds administered in ED - more than 30 mins after primary treatment
Triptan administered in ED -more than 30 mins after primary treatment
Pethidine/Meperidine administered in ED - more than 30 mins after primary treatment
Other Opioid administered in ED - more than 30 mins after primary treatment
Oxycodone administered in ED -more than 30 mins after primary treatment
Chlorpromazine Infusion administered in ED - more than 30 mins after primary treatment
Metoclopramide administered in ED - more than 30 mins after primary treatment
Ondansetron administered in ED - more than 30 mins after primary treatment
Prochlorperazine administered in ED - more than 30 mins after primary treatment
Droperidol/ Haloperidol administered in ED - more than 30 mins after primary treatment
Ergot Alkaloids administered in ED - more than 30 mins after primary treatment
03/17/2020 1:10pm projectredcap.org
Confidential Page 17
Antibiotic/ Antiviral Agent administered in ED - more than 30 mins after primary treatment
Corticosteroid administered in ED - more than 30 mins after primary treatment
Other Medication (oral or parenteral) administered in ED -more than 30 mins after primary treatment
Other medication given > 30 minutes after initial treatment. Provide description __________________________________
Treatment in ED > 30 minutes after initial treatment 1 No 2 Yes
Treatment in ED 1 No Oxygen Therapy 2 Yes
(> 30 minutes after initial treatment )
Treatment in ED 1 No Acupuncture 2 Yes
(> 30 minutes after initial treatment )
Treatment in ED 1 No Intravenous fluids (not part of a drug infusion) 2 Yes
( > 30 minutes after initial treatment )
ED Intubation and mechanical ventilation No Within 30 minutes of arrival at ED After 30 minutes of arrival at ED
Neurosurgical Intervention performed 1 No 2 Yes
Neurosurgical Intervention Time 1 Within 24 hours 2 = >24 hours
Interventional Radiology Performed 1 No 2 Yes
Interventional Radiology Time 1= Within 24 hours 2 = >24 hours
03/17/2020 1:10pm projectredcap.org
Confidential Headache in Emergency Departments (Head Study)
Page 18
Final ED Diagnosis and Disposition
Final ED Diagnosis 1 Primary headache (benign headache not otherwise specified) 2 Migraine 3 Cluster headache 4 Musculoskeletal 5 Tension headache 6 Subarachnoid haemorrhage 7 Other intracranial haemorrhage 8 Post coital headache 9 Neoplasm 10 Viral illness without meningitis 11 Sinusitis 12 Meningitis (viral) 13 Meningitis (bacterial) 14 Meningitis (Fungal) 15 Meningitis(unknown) 16 Encephalitis 17 Stroke 18 Post-traumatic headache 19 Cerebral abscess 20 Toxicity e.g. CO (specify) 21 Trigeminal neuralgia/ cranial neuralgias 22 Glaucoma 23 Alcohol-related hangover 24 Analgesia overuse 25 Temporal arteritis 26 Intracranial hypertension 27 Vascular dissection 28 Shingles (herpes zoster) of head/ neck 29 Other (specify) 30 Unclear
ED Diagnosis (OTHER or TOXICITY) please describe __________________________________
Disposition 1 Home from ED Observation Unit (EOU) 2 Home from ED 3 Admit ward 4 Admit critical care 5 Transfer 6 Unknown 7 Died in ED 8 Theatre 9 Interventional Radiology
03/17/2020 1:10pm projectredcap.org
Confidential Page 19
Final Hospital Diagnosis (for admitted patients only) 1 Primary headache (benign headache not otherwise specified) 2 Migraine 3 Cluster headache 4 Musculoskeletal 5 Tension headache 6 Subarachnoid haemorrhage 7 Other intracranial haemorrhage 8 Post coital headache 9 Neoplasm 10 Viral illness without meningitis 11 Sinusitis 12 Meningitis (viral) 13 Meningitis (bacterial) 14 Meningitis (fungal) 15 Meningitis(unknown) 16 Encephalitis 17 Stroke 18 Post-traumatic headache 19 Cerebral abscess 20 Toxicity e.g. CO (specify) 21 Trigeminal neuralgia/ cranial neuralgias 22 Glaucoma 23 Alcohol-related hangover 24 Analgesia overuse 25 Temporal arteritis 26 Intracranial hypertension 27 Vascular dissection 28 Shingles (herpes zoster) of head/ neck 29 Other (specify) 30 Unclear
(Select from drop down list)
Final Hospital Diagnosis (OTHER or TOXICITY) please describe __________________________________
In-Patient Outcome (for admitted patients only) 1= discharged alive 2= died 3= unknown
(Select from drop down list)
Length of Stay (total days - including day of admission and day of discharge) __________________________________
(Any partial days =1 day. If admitted and discharged within 24 hours = 1 day. )
Medication prescribed at discharge from ED/ ED No Observation Unit Yes
03/17/2020 1:10pm projectredcap.org
Confidential Page 20
No Yes Paracetamol (on discharge from ED or EOU)
Aspirin (on discharge from ED or EOU)
Codeine containing compounds (on discharge from ED or EOU)
NSAID (other than aspirin) (on discharge from ED or EOU)
Triptan (on discharge from ED or EOU)
Oxycodone (on discharge from ED or EOU)
Tramadol (on discharge from ED or EOU)
Other Opioid (on discharge from ED or EOU)
Metoclopramide (on discharge from ED or EOU)
Prochlorperazine (on discharge from ED or EOU)
Ondansetron (on discharge from ED or EOU)
Ergot Alkaloids (on discharge from ED or EOU)
Antibiotic/antiviral agent (on discharge from ED or EOU)
Corticosteroid (on discharge from ED or EOU)
Other medication to treat headache or cause of headache prescibed (on discharge from ED or EOU)
Other ED discharge medications. This refers to medications to treat headache or cause of headache __________________________________
Representation within 72 hours (patients discharged 1 No from ED only) 2 Yes
03/17/2020 1:10pm projectredcap.org
Confidential Page 21
Representation Final ED Diagnosis 1 Primary headache (benign headache not otherwise specified) 2 Migraine 3 Cluster headache 4 Musculoskeletal 5 Tension headache 6 Subarachnoid haemorrhage 7 Other intracranial haemorrhage 8 Post coital headache 9 Neoplasm 10 Viral illness without meningitis 11 Sinusitis 12 Meningitis (viral) 13 Meningitis (bacterial) 14 Meningitis (Fungal) 15 Meningitis(unknown) 16 Encephalitis 17 Stroke 18 Post-traumatic headache 19 Cerebral abscess 20 Toxicity e.g. CO (specify) 21 Trigeminal neuralgia/ cranial neuralgias 22 Glaucoma 23 Alcohol-related hangover 24 Analgesia overuse 25 Temporal arteritis 26 Intracranial hypertension 27 Vascular dissection 28 Shingles (herpes zoster) of head/ neck 29 Other (specify) 30 Unclear
Representation ED Diagnosis (OTHER or TOXICITY) please describe __________________________________
If represented, was patient admitted/ transferred for 1 No admission 2 Yes
Neurosurgery at Representation Visit 1 No 2 Within 24 hours 3 Within 1 week
Interventional Radiology at Representation 1 No 2 Within 24 hours 3 Within 1 week
Supplementary table 1. Sample size by country
Country n (%)
Australia 1,777 (33.6)
Turkey 982 (18.6)
Colombia 757 (14.3)
New Zealand 593 (11.2)
Singapore 579 (10.9)
United Kingdom 276 (5.2)
France 114 (2.2)
Belgium 70 (1.3)
Romania 69 (1.3)
Hong Kong 64 (1.2)
Israel 12 (0.2)
Total 5,293
Supplementary figure 1. Area under ROC curve (AUC) calculated from a multivariate
logistic model of 2137 patients with no missing data for any red flags (papilloedema
was not included as a predictor) – Overall cohort
Supplementary figure 2. Area under ROC curve (AUC) calculated from a multivariate
logistic model of 2137 patients with no missing data for any red flags (papilloedema
was not included as a predictor) – No neurological features cohort
Page updated
Google Sites
Report abuse