內科-1972 Circulation: Effects of Treatment on Morbidity in Hypertension III 年齡-舒張壓, 過去病史. 副作用進一步分析

2026-09-01 10:10AM

Effects of Treatment on Morbidityin Hypertension III. Influence of Age, Diastolic Pressure,and Prior Cardiovascular Disease; 

Further Analysis of Side Effect
Circulation, Volume XLV, May 1972

筆記

1. 降血壓藥物無法減少心肌梗塞與猝死機率(這項有點反直覺)

2. 隨機分組前血壓值: 醫生在隨機分組前最後兩次門診就診期間測量血壓的平均值。

但目前關於血壓測量. 建議使用家中血壓為主(避免白袍症干擾)

3. 50歲以下. 若無心血管風險. 降血壓治療沒有效益. 

註釋與縮寫

1. morbid events 可翻譯為發病事件. 罹病事件. 不良事件 .併發症事件

2. CHF 充血性心衰竭

3. CVR異常: cardio-vascular-renal abnormalities

3. 亞組 subgroup. 將治療組與對照組依照特定條件在

(下面中文使用 GOOGLE 翻譯)
摘要

THE VETERANS Administration Coop-erative Study Group 提供了退伍軍人事務部合作研究的補充數據，該研究納入了194名對照組男性患者和186名治療組男性患者，這些患者的初始舒張壓平均值為90-114 mmHg。

排除條件, 先前有其他高血壓嚴重併發症病史的患者, 包括下列

1. 腦出血或蜘蛛網膜下腔出血

2. 需要持續使用利尿劑的持續性充血性心臟衰竭、

3. 高血壓加速期

4. 急性高血壓腦病變

研究檢視了以下隨機分組前存在的危險因子與疾病發作率和治療效果之間的關係：

(1) 心血管腎臟 (CVR) 異常，其盛行率高於一般高血壓患者族群；

(2) 舒張壓；

(3) 年齡。

疾病發作率 Attack rates 和治療效果 effective-ness of treatment 均與隨機分組前存在的這些危險因子的數量呈正相關。

年齡和入組時是否有CVR異常似乎對後續疾病發作率有顯著影響，而入組血壓水平對疾病發作率的影響相對較小。

另一方面，「治療效果」似乎最受初始血壓水平的影響。

入組前舒張壓在 90 至 104 mmHg 範​​圍內的患者，除非入組時有心血管風險異常或年齡超過 50 歲，否則從治療中獲益甚微。

需要更長的追蹤期來評估治療在低風險亞組的價值。

在副作用方面，治療組輕度低血鉀、高尿酸血症和空腹血糖升高的發生率顯著高於對照組。

應權衡這些及其他副作用與治療低風險高血壓患者預期獲益之間的利弊。

THE VETERANS Administration Coop-erative Study Group 一項隨機、雙盲臨床試驗的結果，該試驗納入了194名對照組男性患者和186名治療組男性患者，初始舒張壓平均為90至114 mmHg，並進行了長達5.5年（平均3.3年）的前瞻性隨訪。 治療藥物為 hydrochlorothiazide, reserpine, hydralazine.

對照組有19例死於心血管疾病，治療組有8例。生命表分析表明，治療使5年內發生致命或非致命性發病事件 morbid events 的風險從55%降低至18%。接受治療的患者中，充血性心臟衰竭、中風和進行性腎損傷的發生率顯著降低或消失。

然而，對照組和治療組的心肌梗塞和猝死發生率基本上相同。除可評估的併發症外，對照組有20位患者出現持續性舒張壓升高（≥125 mmHg），而治療組無一例出現這種情況。

在試驗報告中，患者在隨機分組前心血管異常的盛行率特徵描述不完整。本報告旨在確定既往心血管損傷與治療效果之間的關係。此外，本報告也比原始論文更詳細地分析了年齡和血壓的影響。本報告和初始報告中給出的隨機分組前血壓值代表醫生在隨機分組前最後兩次門診就診期間測量血壓的平均值。

最後，本報告還提供了有關副作用和劑量調整的補充數據。

年齡的影響：

初始報告中所給的中位年齡為：對照組49.2歲，治療組48.1歲。補充數據見表1。

對照組有51%的患者年齡小於50歲，治療組有55%的患者年齡小於50歲。

約四分之一的患者年齡在50-59歲之間。 

60歲以上年齡組佔對照組的22.1%，佔治療組的20.5%，其中年齡最大的患者為75歲。

不同年齡組的隨機分組前血壓如表2。

收縮壓與年齡呈正相關，40歲以下患者的平均收縮壓為154 mmHg，並隨年齡增長而升高，70-75歲年齡組的平均收縮壓為178 mmHg。

然而，各年齡層的平均舒張壓基本上相同。對照組及治療組患者的血壓無顯著差異。(分組前舒張壓)

已知高血壓病程的資料顯示，48%的患者確診高血壓病程為5年或以上，29%的患者確診高血壓病程為9年或以上。高血壓病程較長的患者盛行率相對較高，部分原因可能是年齡分佈，部分原因可能是樣本僅包含「固定型」高血壓患者，即住院第四至第六天舒張壓平均值≥90 mmHg的患者。

表3列出了隨機分組後, 五年研究期間以年齡劃分的併發症 morbid events 發生率。如預期，主要併發症 morbid events 的發生率隨年齡增加而增加。

在對照組中(做為疾病背景值)，50歲以下患者的併發症 morbid events 發生率為15.2%，而59歲以上患者的併發症發生率為62.8%。

在治療組中，這兩個年齡組的併發症 morbid events 發生率分別為6.9%和28.9%。

治療似乎對所有年齡層均有效。

治療預防併發症 morbid events 的有效性是透過比較對照組和治療組患者主要併發症發生率的年齡百分比差異，再除以對照組患者的發生率百分比來估算的（表3）。

在50歲以下的亞組中，治療有效率達55%(15.2% → 6.9%)

在50-59歲年齡組中，治療有效率達68%(26.9%→8.7%

在60歲以上患者中，治療有效率達54%(62.8%→28.9%)

表4顯示了年齡與發病事件類型之間的關係。

在對照組中，年齡最大的患者組最常見的併發症是腦血管意外和充血性心臟衰竭。

充血性心臟衰竭在60歲以下人群中並不常見。

然而，高血壓惡化或進行性腎損傷主要發生於50歲以下的對照組患者。

冠狀動脈疾病在所有年齡層中均有發生，但在60歲及以上患者的發生率較高（表1、4）。

與其他類別不同，冠狀動脈疾病的併發症似乎不受治療的影響。

(這個有點反直覺. 但實驗數據是降壓治療無法降低心肌梗塞或猝死機率)

既往心血管異常的影響：

表5總結了隨機分組前發現的心血管-腎臟異常的盛行率。

在眼底分級時，「高血壓性」和「硬化性」改變是分別評估的。

[註: 以靜脈直徑做對照. 計算動脈直徑與靜脈直徑比例. 動脈越細(狹窄越嚴重). 比值越小. 動脈小於靜脈直徑 50%. 稱為2級改變]

1級高血壓性改變定義為小動脈管徑可能減小，動靜脈比值約為1:2；

2級高血壓性改變則表示小動脈管徑明顯減小，動靜脈比值小於1:2，甚至細如絲線。

由於1級改變與正常變化難以區分且變異性較大，因此僅報告2級變化。

此類改變在31%的對照組患者和25%的治療組患者中出現。

 2級硬化性改變（而非高血壓性改變）定義為散在的顯著動靜脈缺損或小動脈節段不對稱不規則。

此類改變在24%的患者中出現。

分析了背景特徵以確定是否存在心臟、中樞神經系統或腎臟異常。

所有在上述任何目標器官系統中，嚴重程度評分大於零的患者均被納入研究，但以下情況除外：頭痛（如果這是唯一的CNS異常）和劇烈活動後呼吸困難（如果這是唯一的心臟異常）；眼底評分也不納入研究。

納入標準包括：

日常活動呼吸困難、心絞痛、左心室擴大（LVE）、X光片顯示心臟擴大、除頭痛以外的其他神經系統症狀或腎臟評分大於零。

根據這些標準，對照組中有55%的患者和治療組中有60%的患者在主要目標器官系統中表現出一種或多種異常（表5）。

採用Ungerleider標準，透過標準胸部後前位X光片確定心臟擴大。

根據這些標準，隨機分組前，對照組中有 22% 的患者和治療組中有 28% 的患者被認為有心臟擴大。

左心室擴大 (LVE) 的心電圖標準要求患者同時出現電壓變化（V1 或 V2 導聯的 S 波和 V3 或 V4 導聯的 R 波 > 35 mm）以及 I、aVI 和 V1 或 V2* 導聯的 T 波平坦、雙相或倒置。

在隨機分組前，對照組和治療組中各有 16% 的患者符合這些 LVE 標準。

腎損傷分級如下。

1級腎臟病變包括以下任兩項：

三次獨立過夜尿液收集的比重均≤1.020；

任一尿液樣本中蛋白尿≥1+；

以及2小時混合尿液樣本中酚磺酞（PSP）排泄率≤45%。

2級腎臟病變包括以下任兩項：

比重≤1.015；三次每日尿液樣本中蛋白尿≥1+；

以及PSP排泄率≤30%。 

3級腎臟病變指上述三項全部符合。 

4級腎臟病變指有氮質血症。

對照組和治療組中1級和2級腎臟病變的發生率相同，分別為12%和2%（表5）。

所有患者均未出現3級腎臟病變。

對照組有4例患者，治療組有3例患者的血中尿素氮濃度在25至32 mg/100 ml範圍內。然而，這4例患者的血清肌酸酐值正常，且在隨機分組後未發生不良事件。

其餘3例患者中，1例死於膀胱癌，1例發生充血性心臟衰竭，另1例最初患有慢性腎絲球腎炎（GN），血清肌酸酐為2.4 mg/1100 ml，後因舒張壓升高而退出試驗。 

5例患者（對照組2例，治療組3例）被診斷為原發性腎臟疾病。

其中4例診斷為慢性腎絲球腎炎，1例診斷為雙側髓質海綿腎。

兩名慢性腎絲球腎炎對照組患者中，一名因舒張壓升高（見上文）而退出研究，另一名患者則發生腦血栓。

三名接受積極治療的原發性腎病患者均未發生不良事件。

研究納入了一些先前發生過重大事件的患者。 

380名患者中，7%（對照組和治療組各佔一半）在隨機分組前曾發生心肌梗塞。

對照組有6%的患者，治療組有9%的患者曾經有心臟衰竭病史，但在隨機分組時未出現充血性心臟衰竭。

對照組和治療組中各有5%的患者被臨床診斷為腦血栓。

先前有其他高血壓嚴重併發症病史的患者，例如腦出血或蜘蛛網膜下腔出血、需要持續使用利尿劑的持續性充血性心臟衰竭、高血壓加速期或急性高血壓腦病變，均未納入本試驗。

老年患者心血管腎臟異常的發生率較高（表6）。 

50歲以下患者中，46%出現一種或多種異常，而50-59歲年齡組的比例為65%，59歲以上患者中這一比例為78%。

既往心血管損傷顯著增加了隨機分組後發生不良事件的風險（表7）。

在隨機分組前已發生心肌梗塞、充血性心臟衰竭或腦血栓的患者中，隨機分組後，對照組和治療組的後續主要併發症發生率分別為53%和26%。

在隨機分組前無重大併發症但有心臟、中樞神經或腎臟損傷的患者中，對照組後續不良事件發生率為33%，治療組為8%。對於隨機分組前有異常的合併亞組，治療有效率達64%（表7）。

在無異常的對照組患者中，不良事件發生率較低。

在本子樣本中，對照組患者發生嚴重併發症的比例為16%，而治療組為8%。儘管差異無統計學意義，但治療有效率達50%的趨勢與先前有異常的患者組相似。

需要強調的是，20名對照組患者在發生任何不良事件之前就因舒張壓升高至124 mmHg以上且持續3週或更長時間而被排除出試驗。 這20名患者中有7名在隨機分組前的門診舒張壓低於105 mmHg。 20名患者中有15名年齡小於50歲；20名患者中有10名沒有心臟、中樞神經或腎臟異常的證據。由於在這種舒張壓水平下，若不進行治療，發生後續事件的風險極高⁴，因此在發生不良事件之前將這些患者排除在外，可能導致所有亞組（尤其是50歲以下亞組和無心臟、中樞神經系統或腎臟異常的亞組）的治療效果被低估。在隨機試驗期間發生的27例死亡病例中，17例死亡與心肌梗塞相關或猝死；其中11例發生在對照組，6例發生在治療組。關於這17例患者的其他危險因素，對照組中有6例，治療組中有2例在隨機分組前血清膽固醇高於260 mg/100 ml。對照組有2位患者，治療組有1位患者的空腹血糖值高於110 μg/100 ml。在所有猝死患者中，年度檢查均未發現低血鉀的證據。 11名對照組患者出現充血性心臟衰竭。其中5例為復發，6例為首次發作，發生在隨機分組後。儘管17名治療組患者在隨機分組前有充血性心臟衰竭病史，但在隨機分組後均未復發。

影響治療效果的因素：

由於舒張壓程度、年齡、考慮到先前心血管異常對發病事件發生率的影響，評估這些因素各種組合的影響顯得特別重要。將患者群體細分為若干小亞組會增加觀察率因隨機波動而出現較大誤差的可能性。因此，我們採用多元迴歸技術來估計發病率，該技術類似於曲線擬合，可以平滑觀察資料中的一些隨機波動。多元迴歸技術同樣可以為較大的亞組（所有具有特定危險因子的患者）提供更精確的估計值。因此，該技術已被用於重新計算先前描述的亞組的發病率，例如，年輕組與老年組。表 8 第一部分列出了這些改進後的估計值。當然，計算出的「治療效果」在某種程度上有所修正。就風險因素組合而言，結果表明，無論是否存在明顯的心血管腎臟疾病，也無論年齡大小，舒張壓在 105-114 mmHg 的亞組中，治療獲益最大（表 8，圖 1）。

治療效果最差的兩個亞組是：

初始舒張壓低於 105 mmHg 且無心血管、中樞神經系統或

腎臟異常的患者；以及 50 歲以下且血壓水平相似的患者。

當然，這些結果可能受到相對較短的追蹤期的顯著影響。表 9 顯示了三種風險因子的整體盛行率與發病率和治療效果之間的關係，而與風險因子的類型無關。當不存在危險因子時，疾病發作率和治療效果均最低；隨著危險因子盛行率的增加，發作率和治療效果也逐漸提高。

副作用：

初步報告顯示，兩名接受治療的患者因疑似毒性反應而退出試驗，其中一名患者出現過敏性紫斑，另一名患者疑似患有狼瘡綜合徵，但尚未確診。此外，還有30名患者因疑似藥物相關副作用而停用了疑似致病藥物或其安慰劑。其中12名患者出現憂鬱症狀，7名在治療組，5名在對照組。 10名患者出現消化性潰瘍；其中6名服用活性藥物，4名服用安慰劑。兩名患者（一名服用活性藥物，另一名服用安慰劑）因陽痿而改變了治療方案。其餘6名患者均接受活性藥物治療。

其副作用包括嗜睡、嚴重鼻塞、痛風、可能由低血壓引起的癲癇發作以及葡萄糖耐受試驗異常。

生化副作用：低血鉀症。如表10所示，隨機分組前，對照組和治療組患者的血清鉀濃度分佈幾乎相同，對照組和治療組分別只有1%和2%的患者血清鉀濃度低於正常值。然而，在第一次年度檢查時，治療組的血清鉀水平百分比分佈向低值偏移，而對照組基本上保持不變。 1年後，對照組和治療組分別有2%和23%的患者血清鉀濃度低於正常值。然而，治療組中僅有1%的患者血清鉀濃度低於2.5 mEq/L。第二次年度檢查也觀察到了類似的趨勢，儘管治療組中沒有患者的血清鉀水平低於2.5 mEq/L。在完成兩次年度檢查的患者中，在第一次年度檢查時血清鉀水平低於3.5 mEq/L的23名治療組患者中，約有一半在第二次年度檢查時仍存在低鉀血症。在首次年度檢查中，5名血清鉀值低於正常值的患者接受了補鉀治療。其中3名患者在第二次年度檢查中血清鉀值恢復正常，而2名患者仍存在輕度低血鉀症。

血清尿酸：

初次檢查時，對照組13%的患者和治療組11%的患者尿酸值達到或超過8.0 mg/100 ml（表11）。在首次年度檢查中，尿酸水平達到或超過8.0 mg/100 ml的患者在對照組中佔16%，在治療組中佔30%；在第二次年度檢查中，對照組和治療組的比例分別為19%和25%。在40名治療一年後出現高尿酸血症的患者中，有18名患者在第二次年度檢查時尿酸水平仍達到或超過8.0 mg/100 ml。在首次年度檢查中，5名尿酸水平升高的患者分別接受了別嘌醇或丙磺舒治療；在第二次年度檢查中，3名患者的尿酸水平仍然升高，2名患者的尿酸水平降至8.0mg/100ml以下。如前所述，2名接受積極治療的患者在隨機分組後出現了首次痛風發作。

在首次年度檢查時血清鉀濃度低於 3.5 mEq/L 的 23 例接受治療的患者中，約有一半在第二次年度檢查時仍有低血鉀。首次年度檢查時血清鉀水平低於正常值的 5 例患者接受了鉀補充劑治療。其中 3 例在第二次年度檢查時血清鉀濃度恢復正常，而 2 例仍有輕度低血鉀。血清尿酸：在初次檢查時，對照組中有 13% 的患者和治療組中有 11% 的患者尿酸濃度達到或超過 8.0 mg/100 ml（表 11）。在首次年度檢查時，尿酸水平達到或超過 8.0 mg/100 ml 的患者在對照組中佔 16%，在治療組中佔 30%；在第二次年度檢查時，對照組和治療組的相應比例分別為 19% 和 25%。在接受治療一年後出現高尿酸血症的40例患者中，18例在第二次年度檢查時尿酸水準仍為8.0 mg/100 ml或更高。在第一次年度檢查時尿酸升高的5例患者分別接受了別嘌醇或丙磺舒治療；在第二次年度檢查時，3例患者的尿酸水平仍然升高，2例患者的尿酸水平降至8.0 mg/100 ml以下。如前所述，2例接受積極治療的患者在隨機分組後出現了首次痛風發作。

空腹血糖。

由於在試驗的前3年中，部分醫院以餐後2小時血糖取代空腹血糖，因此空腹血糖（FBS）數據不完整。

在隨機分組前，對照組和治療組中空腹血糖（FBS）值≥110 mg/100 ml的盛行率基本相同（表12）。然而，隨機分組一年後，治療組的盛行率為20.8%，而對照組為15.6%；隨機分組兩年後，治療組的盛行率為30.0%，對照組為17.0%。在隨機分組前，治療組中有7.0%的患者FBS水平≥120 mg/100 ml；在第一次年度檢查時，這一比例為12.5%；在第二次年度檢查時，這一比例為14.6%。由於第一年和第二年追蹤結束時可供檢查的患者總數存在差異（表12），因此，數據分析僅使用完成了第一次和第二次年度檢查的患者作為分母。結果仍顯示，治療組中空腹血糖升高的患者比例高於對照組患者。治療組患者血清鉀降低與空腹血糖升高無顯著相關性。治療一年後，119例空腹血糖低於110 mg/100 ml的治療組患者平均血清鉀為3.91 mEq/L（標準差0.57 mEq/L），40例空腹血糖為110 mg/100 ml或更高的治療組患者平均血清鉀為3.89 m兩年後，兩組患者的血清鉀濃度分別為 4.04 ± 0.58 mEq/L 和 3.98 ± 0.91 mEq/L。

方案藥物劑量調整：

隨機分組時，治療組所有患者均接受氫氯噻嗪 50 mg 和利血平 0.1 mg 每日兩次合併肼屈嗪 25 mg 每日三次的治療方案。 * 顯然，與通常使用單一降血壓藥物的初始治療方案相比，這種治療方案會導致更高的副作用發生率。因此，當出現低血壓症狀或其他副作用時，可以減少劑量或停用三種藥物中的一種。從隨機分組的 380 名患者中抽取了 124 名患者作為樣本，檢查了他們的病歷，以確定劑量調整情況及其原因。樣本的選擇方式確保每個參與診所的樣本數量比例均衡，並涵蓋患者入組試驗的時間段。對照組中有 9 位（16%）患者調整了劑量，治療組中有 36 位（53%）患者調整了劑量（表 13）。

安慰劑組調整劑量的原因包括以下副作用：

頭痛、鼻塞、心絞痛、皮疹和多種不適症狀。治療組 36 例調整劑量的患者中有 28 例，其劑量減少的主要原因是出現低血壓症狀，如虛弱、嗜睡或暈厥，這些症狀通常伴隨舒張壓降至遠低於 90 mmHg 的水平。

在適當降低劑量後，這些症狀均有所緩解。

除低血壓症狀外，其他導致劑量調整的副作用包括 3 例心絞痛、3 例頭痛、一名患者出現緊張，另一名患者出現鼻塞。

因此，除低血壓導致劑量調整外，其他副作用的發生率在對照組和治療組之間無顯著差異。

如果治療開始時僅使用一種降血壓藥，低血壓症狀可能會少很多。

主觀副作用的發生率：

對同一組124名患者的主觀副作用進行了評估，評估依據是門診就診報告表上的醫生訪談清單。回顧了隨機分組前試驗期間最後兩次就診的報告以及所有隨機分組後的就診報告。可能由於不同門診醫師的價值判斷不同，不同醫院報告特定症狀的頻率有相當大的差異。由於每位患者始終由同一位醫生接診，因此將每位患者的隨機分組後時期與隨機分組前時期進行比較是合理的。對於任何特定患者，僅統計隨機分組後出現且在隨機分組前未記錄的副作用。

結果顯示，在隨機分配至安慰劑組的患者中，有相當數量的患者報告了特定的副作用。

例如，在隨機分組前未報告惡夢的52名對照組患者中，有6名在隨機分組後的某個時間報告了這種「副作用」（表14）。

對於惡夢、關節炎、心絞痛​​和頭痛這些特定症狀，對照組患者的發生率實際上高於治療組。

而對於憂鬱、皮疹、陽痿和其他「不適」等副作用，對照組和治療組的發生率幾乎相同。

對「副作用」發生率高的一個可能解釋是，醫生在隨機分組後可能更重視副作用訪談。

治療組患者發生頻率較高的副作用僅有嗜睡或疲倦、鼻塞、潰瘍症狀及首次出現任何不適症狀。

後者指的是隨機分組前沒有任何不適症狀，但在隨機分組後報告出現一種或多種副作用的患者。

此外，在達到適當的降血壓藥物維持劑量後，檢查副作用的發生率也十分重要。因此，表14中也列出了特定副作用的報告數據，但省略了隨機分組後前兩次訪視（僅包括第三次及以後的訪視）。

在這些後續訪視中，對照組和治療組患者報告的潰瘍症狀、嗜睡或疲勞以及鼻塞症狀的發生率差異無統計學意義。

值得注意的是，在這些追蹤期間，對照組患者報告心絞痛和頭痛的比例明顯高於治療組患者。

討論：

本研究中併發症的發生率高於高血壓患者整體族群的預期水準。原因可能包括：（1）入院時患有高血壓但住院第四至第六天舒張壓平均值低於90 mmHg的患者被排除在外；（2）超過一半的患者出現心血管或腎臟異常；（3）近30%的患者已知患有高血壓10年或更長時間。 (4) 雖然只有五分之一的患者年齡超過 60 歲，但這些患者卻佔了所有發病事件的一半。對照組中，初始舒張壓平均在 105-114 mmHg 的患者發生事件發生率高於初始舒張壓在 90-104 mmHg 的患者。治療效果在初始血壓較高的患者中更為顯著。本研究表明，既往心血管疾病顯著增加對照組發生事件的風險，並且在有限的觀察期內，有既往疾病史的患者治療效果更佳。由於治療組與對照組相比獲益明顯，因此必須終止試驗。由於無心血管疾病史的患者風險較低，因此追蹤時間過短，不足以觀察到大量事件，也無法在該亞組樣本中獲得對照組和治療組之間具有統計意義的差異。

然而，在「無異常」組中觀察到的50%的治療有效率與「異常」組中64%的有效率並無顯著差異，這與治療的保護作用相符。

本試驗的結果表明，有必要加大力度識別並持續為出現任何心血管損傷跡像或舒張壓平均值超過104 mmHg的患者提供充分的治療。

來自代表性人群調查的證據表明，許多高血壓患者要么不知道自己患有高血壓，要么沒有接受充分的治療。

然而，還需要更多證據來確定治療的益處是否大於其對低風險患者（例如輕度高血壓且無血管疾病證據的患者，特別是女性患者以及血壓波動性高血壓患者）的弊端。

目前所有可用的抗高血壓藥物都可能引起毒性反應，而副作用，特別是與噻嗪類及其相關利尿劑相關的生化改變，較為常見。

由於其他令人不安的副作用，通常需要調整劑量。雖然這些副作用的風險看似很小，但必須與預期獲益進行比較，尤其是在高血壓併發症風險也相對較低的患者中。然而，如果患者未接受治療，則應定期隨訪，以確定高血壓是否進展至更嚴重的階段。這種追蹤對於年輕患者尤其重要。

在本研究中，20名舒張壓嚴重升高的對照組患者中有15名年齡在50歲以下；其中7名患者的初始舒張壓低於105 mmHg。

在本研究的患者樣本中，抗高血壓治療似乎能有效降低高血壓相關併發症，但心肌梗塞和猝死除外。

這項結果並非一定與統計學證據相矛盾，即高血壓是冠心病發病率增加的「危險因子」之一。更大的樣本量或更長的追蹤期或許能夠揭示本研究中未發現的差異。此外，如果在高血壓早期階段就開始治療，或許能夠獲得更大的保護作用。為了解決這個問題，需要在不同的高血壓患者群體中進行治療試驗。需要對不同高血壓患者族群進行治療試驗，以解決這個問題。需要對不同高血壓患者族群進行治療試驗，以解決這個問題。

SUMMARYA

Additional data are presented from the Veterans Administration Cooperative Studywith respect to the 194 control and 186 treated male patients with initial diastolicblood pressures averaging 90-114 mm Hg. Attack rates and effectiveness of treatment were examined with respect to the following risk factors present at entry: 

(1) cardio-vascular-renal (CVR) abnormalities, the prevalence of which was higher than in thegeneral population of hypertensive patients; 

(2) diastolic blood pressure; and 

(3) age.Both attack rates and effectiveness of treatment increased directly with the number ofthese risk factors present at entry. 

Age and presence of CVR abnormalities at entryappeared to strongly influence subsequent attack rates, whereas entry level of bloodpressure had a relatively smaller effect on attack rates. On the other hand, "effective-ness of treatment" appeared to be most influenced by the initial level of blood pressure. Patients with prerandomization diastolic blood pressure in the range of 90 to104 mm Hg derived relatively little benefit from treatment unless they had CVR abnormalities at entry or were over 50 years of age. A longer period of follow-up wouldbe needed to assess the value of treatment in the lower risk subgroups.With respect to side effects, the incidence of mild hypokalemia, hyperuricemia, andelevated fasting blood sugar was significantly higher in the treated group. These andother side effects should be weighed against the benefit to be expected from treatinghypertensive patients at low risk.

THE VETERANS Administration Coop-erative Study Group previously reportedon the results of a randomized, double-blindclinical trial in 194 control and 186 treatedmale patients with initial diastolic bloodpressures averaging 90 through 114 mm Hgfollowed prospectively for periods up to 5.5years, average 3.3 years.1 Treatment consistedof a combination of hydrochlorothiazide,reserpine, and hydralazine. There were 19deaths related to cardiovascular disease in thecontrol group and eight in the treated series.Life table analysis indicated that the risk of amorbid event, fatal or nonfatal, over a 5-yearperiod was reduced from 55 to 18% bytreatment. Congestive heart failure, stroke,and progressive renal damage were sharply reduced or eliminated in the treated patients.However, the incidence of myocardial infarc-tion and sudden death was essentially thesame in the control and treated groups. Inaddition to assessable morbid events, 20 control patients versus none of the treatedgroup developed persistent diastolic elevationsof 125 mm Hg or higher.

In the trial report the patients wereincompletely characterized as to the preva-lence of cardiovascular abnormalities prior torandomization. The present report determinesthe relationship between prior cardiovasculardamage and the effectiveness of treatment.The influences of age and blood pressure alsoare analyzed in more detail than in theoriginal paper. The prerandomization bloodpressures given in this and in the initialreport1 represent the average of the readingstaken by the physician during the last twooutpatient visits preceding randomization.Finally, additional data with respect to sideeffect and dose modifications are presented.Influence of AgeThe median ages were given in the initialreport1 as 49.2 years for the control group and48.1 years for the treated series.Additional data are presented in table 1.Fifty-one percent of the control and 55% of thetreated group were less than 50 years of age.Approximately one fourth of the patients were in the 50-59-year age group. The 60 and aboveage group included 22.1% of the control and 20.5% of the treated patients, the oldest patientbeing 75 years of age.The prerandomization blood pressures forthe different age groups are shown in table 2.Systolic blood pressure was related directly toage, and averaged 154 mm Hg in the patientsunder 40 years of age and rose with age to 178mm Hg in the 70-75-year age group. Meandiastolic blood pressures, however, were es-sentially the same at all ages. There were nosignificant differences in blood pressure in thecontrol and treated patients.Information on the duration of knownhypertension indicated that 48% had recog-nized hypertension for 5 years or more, and29% had hypertension for 9 years or more. Therelatively high prevalence of hypertension oflong duration may be explained in part by theage distribution and in part by the fact thatthe sample included only patients with "fixed"hypertension, namely those with diastolicblood pressures averaging 90 mm Hg or morefrom the fourth through the sixth day ofhospitalization.The incidence of morbid events during thestudy by age at randomization is listed intable 3. As would be expected, the incidenceof major complications rose with age. In the control group, 15.2% of the patients under age 50 years developed morbid events followingrandomization as compared to 62.8% of thepatients above age 59 years. In the treatedpatients, the percentage incidence of morbidevents was 6.9 and 28.9%, respectively, in thesetwo age groups. Treatment appeared to beeffective in all age groups. The effectiveness of treatment in preventing morbid events wasestimated from the difference in the percent-age incidence of major complications between control and treated patients divided by thepercentage incidence in the control patients(table 3). 

Treatment was 55% effective in the subsample below 50 years of age, 68% in the50-59-year age group, and 54% in the patientsaged 60 years and over.The relationship between age and type ofmorbid event is shown in table 4. In thecontrol group the most frequent complicationsin the oldest age group were cerebrovascularaccidents and congestive heart failure. Thelatter was uncommon below age 60 years.However, increasing hypertension or progres-sive renal damage occurred predominantly inthe control patients below age 50 years.Coronary artery disease occurred in all agegroups but its percentage incidence was morecommon in the patients aged 60 years or over(tables 1, 4). Unlike the other categories, thecomplications of coronary artery disease didnot appear to be affected by treatment.Influence of Prior CardiovascularAbnormalitiesThe prevalence of cardiovascular-renal ab-normalities found prior to randomization issummarized in table 5. In grading the opticfundi, "hypertensive" and "sclerotic" changeswere evaluated separately. Grade 1 hyperten-sive changes were defined as a probabledecrease in arteriolar caliber with an A-V ratioof approximately 1:2, while grade 2 changesindicated a definite decrease in arteriolarcaliber ranging from an A-V ratio less than 1:2to threadlike arterioles. Because of the difficul ty and variability in differentiating grade 1from niormal, only the grade 2 changes arereported. Such changes occurred in 31% of thecontrol and 25% of the treated patients. Grade2 sclerotic as opposed to hypertensive changeswere defined as scattered prominent A-V nicksor asymmetric irregularity of arteriolar seg-ments. Such changes were reported in 24% ofpatients.Backgound charactistics were analyzed todetermine the presence of cardiac, centralnervous system (CNS), or renal abnormal-ities. All patients with severity grades2 greaterthan zero with respect to any of these targetorgan systems were included with the follow-ing exceptions: headache, if this was the onlyCNS abnormality, and dyspnea on heavyeffort if it was the only cardiac abnormality; optic fundi scores also were not included.Patients with dyspnea on ordinary activity,angina, left ventricular enlargement (LVE),cardiomegaly by X-ray, neurologic symptomsother than headache, or renal score greaterthan zero were included. On the basis of thesecriteria, 55% of the control group and 60% ofthe treated patients exhibited one or moreabnormalities in the major target organsystems (table 5).Ungerleider criteria3 were used for deter-mining cardiomegaly from the standard poste-rior-anterior X-ray of the chest. By thesecriteria, 22% of the control group and 28% of the treated patients were considered to exhibitcardiomegaly prior to randomization.The electrocardiographic criteria for leftventricular enlargemenit (LVE) required thatthe patieint exhibit both voltage changes (S inV, or V, plus R in V, or V,, > 35 mm) and flat,biphasic, or negative T waves in leads I, aVI,and V, or V*;. Sixteeni percent of both thecontrol and treated groups exhibited LVE bythese criteria prior to randomization.Renal damage was graded as follows. Grade1 inicluded any two of the following: specificgravity of 1.020 or less in all of three separateovernight urine collections, proteinuria of 1+or more in any one of these speceimens, andphenosulfonphthalein (PSP) excretion of lessthan 45% in a pooled 2-hour specimen. Grade 2changes included any two of the following:specific gravity of 1.015 or less, proteinuria 1+or more in all of three daily specimens, andPSP excretion of 30% or less. Grade 3 indicatedall three of the above changes. Grade 4denoted the presence of azotemia.The frequency of grades 1 and 2 renalchanges was the same in control and treatedgroups with 12% exhibiting grade 1 and 2%exhibiting grade 2 changes (table 5). None ofthe patients exhibited grade 3 changes. Therewere four patients in the control group andthree in the treated series in whom the bloodurea nitrogen was reported as being in therange of 25 to 32 mg/ 100 ml. However, theserum creatinine value was normal in four ofthese patients, and they did not developmorbid events following randomization. Ofthe remaining three patients, one died of acarcinoma of the urinary bladder, anotherdeveloped congestive heart failure, while thethird, who had chronic glomerulonephritis(GN) with serum creatinine of 2.4 mg/1100 mlinitially, was later removed from the trialbecause of elevated diastolic blood pressure.Five patients, two in the control and threein the treated group, were diagnosed ashaving primary renal disease. Chronic glomer-ulonephritis was diagnosed in four andbilateral medullary sponge kidney in one. Oneof the two control patients with chronic GNwas removed because of elevated diastolic blood pressure (see above) while the otherdeveloped a cerebral thrombosis. In the threeactively treated patients with primary renaldisease there were no morbid events.Included in the study were some patientswho had previous major events. Seven percentof the 380 patienits, equally divided betweencontrol and treated groups, had sustained amyocardial infaretion prior to randomization.Six pereent of the control group and 9% of thetreated patients had a past history of cardiacdecompensation but were Inot in conigestiveheart failure at the time of randomizationi.Five percent of patients in both the controlanid treated groups had a clinical diagniosis ofcerebral thrombosis. Patients with a histoiy ofother major complications of hypertensionisuch as cerebral or subarachnoid hemorrhage,persistent congestive heart failure requiringcontiniuous diuretics, accelerated phase ofhypertensioni, or acute hypertensive enicepha-lopathy were not admitted into the trial.Cardiovascular-renal abnormalities weremore frequent in the older patients (table 6).In those under 50 years of age, 46% presentedwith one or more abnormalities as opposed to65% of the 50-59-year age group and 78,/ of thepatienits above age 59 years.The presence of prior cardiovascular dam-age greatly increased the risk of dexelopingnmorbid events followinig randomizationi (table]7). In the patients with either myocardialinfaretion, congestive heart failure, or cerebralthrombosis prior to ranidomization, subsequentmajor complications occurred in 53% of thecontrol and 26% of the treated group duringthe postrandomization period. In those witlevidence of cardiac, CNS, or renal damagebut without a major complication precedingranidomizationi, the incidenice of subsequentmorbid events was 33% in the control patientsanid 8% in the treated group. Effectixeniess oftreatment for the combined subsamples pre-senting with prerandomnization abnormnalitieswvas 64% (table 7).The incidence of morbid events xwas muehless in the control series of patients presenitingwithout abnormalities. In this subsample, 16%of the colntrol patienlts developed a major complication as opposed to 8% of the treated.The difference is not statistically significantalthough the trend indicating 50% effectivenessof treatment is similar to that found in thegroup with preexisting abnormalities.It should be emphasized that 20 controlpatients were removed from the trial prior toany morbid event because of elevations ofdiastolic pressures to more than 124 mm Hgwhich persisted for 3 weeks or longer.1 Sevenof the 20 patients had a prerandomizationclinic diastolic blood pressure lower than 105mm Hg. Fifteen of the 20 were less than 50years of age; 10 of the 20 had no evidence ofcardiac, CNS, or renal abnormality. Since atthis level of diastolic blood pressure the risk ofdeveloping subsequent events without treat-ment is very high,4 the removal of thesepatients prior to the development of a morbidevent probably resulted in an underestimateof the effectiveness of treatment in allsubgroups but especially in the subgroupunder age 50 years and the group presentingwithout cardiac, CNS, or renal abnormalities.Among the 27 fatalities occurring during therandomized trial, 17 deaths were associated with myocardial inifarction or occurred sud-denly; of these, 11 occurred in the controlgroup and six in the treated patients. Withrespect to other risk factors in these 17patients, six of the control and two of thetreated patients had prerandomization serumcholesterol levels greater than 260 mg/100 ml.Two control patienits and one treated patientexhibited fasting blood sugar levels above 110iig/ 100 ml. There was no evidence ofhypokalemia during the annual examinationsin any of the patients who had suddendeath.Congestive heart failure occurred in 11 ofthe control patients. In five it represented arecurrence, while in six the initial attackoccurred following randomization. Although17 of the treated patients had a history ofcongestive heart failure prior to randomiza-tion, none developed recurrences during thepostrandomization period.Influences on Therapeutic EffectivenessBecause of the separate effects of level ofdiastolic blood pressure,' age, and prior car-diovascular abnormalities on the incidence morbid events it seemed of initerest to assessthe influence of various combinations of thesefactors. Subdivision of the patient populationinto small subgroups increases the likelihoodof large errors in observed rates due torandom fluctuations. Therefore, rates wereestimated by the use of multiple regressiontechnics which, like curve fitting, smooth outsome of the random fluctuations in theobserved data.The multiple regression technic likewise canprovide improved estimates for the large subgroups (all patients presenting with agiven risk factor). Thus, the technic has beenused to restate the attack rates for subgroupspreviously described, e.g., the younger groupversus the older group. Such improved esti-mates are shown in the first section of table 8.Of course, the calculated "effectiveeness oftreatment" is modified to some degree.With regard to combinations of risk factorsthe results indicate that the greatest benefit oftreatment was achieved in the subgroups withdiastolic blood pressure of 105-114 mm Hg regardless of evident cardiovascular-renal dis-ease and irrespective of age (table 8, fig. 1).Treatment was least effective in two sub-groups: those with initial diastolic bloodpressure below 105 mm Hg and no cardiovas-cular, CNS, or renal abnormalities; and thosebelow age 50 years with similar levels of bloodpressure. These results, of course, may beconsiderably influenced by the relatively briefperiod of follow-up.Table 9 indicates attack rates and effective-ness of treatment in relation to prevalenice ofthe three risk factors at entiy irrespective of the type of risk factor. Both attack rates andeffectiveness of treatment are lowest when norisk factors are present anid increase progres-sively with increasing prevalence of riskfactors.Side EffectsThe initial report' indicated that twotreated patients w,ere removed from the trialbecause of presumed toxic reactioiis onebeing aniaphylactoid purpura. the other sus-pected, but unproveni, lupus syndrome. Inadditioni, there wvere 30 patienits in whom the suspected offending drug or its placebo wasdiscontinued because of presumed drug-relat-ed side effects. Twelve of these patientsdeveloped mental depression, of whom sevenwere in the treated group and five in thecontrol group. Ten patients developed pepticulcer; six had been taking active drugs andfour placebos. Two patients, one on activedrugs the other on placebos, had a change inregimen because of impotence. The remainingsix patients all were receiving active treat-ment. Their side effects included excessivesleepiness, severe nasal stuffiness, gout, sei-zures presumably caused by hypotension, andan abnormal glucose tolerance test.Biochemical Side EffectsHypokalemia. As indicated in table 10, thedistribution of serum potassium levels prior torandomization was nearly identical in thecontrol and treated patients with only 1% ofthe control and 2% of the treated patientsexhibiting subnormal values. At the firstannual examination, however, the percentdistributions in the treated group shiftedtoward lower values of serum K while thecontrol group remained essentially unchanged.At the end of 1 year, 2% of the control and 23%of the treated group exhibited subnormalvalues. However, only 1% of the treatedpatients exhibited a serum K level below 2.5mEq/liter. A similar trend was found at thesecond annual examination, although therewere no patients in the treated group who hada serum potassium level below 2.5 mEq/liter.Of the patients completing two annualexaminations, approximately half of the 23 treated patients who exhibited serum K levelsunder 3.5 mEq/liter at the first annualexamination remained hypokalemic at thesecond annual examination. Five of thepatients exhibiting subnormal serum K valuesat the first annual examination receivedpotassium supplements. Three of these exhib-ited normal serum K values at the secondannual examination, while two remainedmildly hypokalemic.Serum Uric Acid. At the initial examination,13% of the control group and 11% of thetreated patients exhibited uric acid levels of8.0 mg/100 ml or higher (table 11). At thefirst annual examination patients with eleva-tions of 8.0 mg/ 100 ml or higher accounted for16% in the control and 30% in the treated series;at the second annual examination the percent-ages were 19 in the control and 25 in thetreated. In the 40 treated patients exhibitinghyperuricemia after 1 year, 18 had uric acidlevels remaining at 8.0 mg/ 100 ml or higher atthe second annual examination. Five of thepatients exhibiting uric acid elevations at thefirst annual examination were placed on eitherallopurinol or probenecid; the uric acidremained elevated in three and fell below 8.0mg/ 100 ml in two at the second annualexamination. As indicated in the previousreport,' two actively treated patients experi-enced their first attack of gout during thepostrandomization period.Fasting Blood Sugar. The data with regardto fasting blood sugar (FBS) are incompletedue to the fact that the 2-hour postprandialblood sugar was substituted for FBS at a fewhospitals during the first 3 years of the trial.


The prevalence of FBS values of 110 mg/100ml or higher was essentially the same in thecontrol and treated groups prior to randomiza-tion (table 12). One year after randomization,however, the prevalence was 20.8% in thetreated series as compared to 15.6% in thecontrol, and 2 years after randomization it was30.0% in the treated and 17.0% in the controlpatients. FBS levels of 120 mg/100 ml werefound in 7.0% of the treated patients prior torandomization, 12.5% at the first annual, and14.6% at the second annual examination.Because of the difference in total patientsavailable for examination at the end of thefirst and second year of follow-up (table 12),the data also were analyzed using as denom-inators only the patients who completed boththe first and second annual examinations. Theresults still indicated a greater percentage ofpatients with elevated FBS in the treated ascompared to the control group of patients.There was no significant correlation be-tween reduction in serum K and the increase in FBS in the treated group of patients. Atter1 year of treatment mean serum K was 3.91 SD0.57 mEq/liter in 119 treated patients withFBS levels below 110 mg/100 ml and 3.89 SD0.57 mEq/liter in 40 treated patients with FBSlevels of 110 mg/100 ml or higher. After 2years, the serum K levels were 4.04 SD 0.58 and3.98 SD 0.91 mEq/liter, respectively, in the twogroups.Changes in Doses of Protocol DrugsAt the time of randomization a combinationof hydrochlorothiazide 50 mg and reserpine0.1 mg twice daily plus hydralazine 25 mgthree times daily5 were administered to allpatients in the treated group.* Obviously, thistherapeutic regimen will result in a higherincidence of side effects than the usual methodof initiating treatment with a single antihyper-tensive agent. Therefore, provision was made for reducing doses or for discontinuing one ofthe three drugs in the presence of hypotensivesymptoms or other side effects.A sample of 124 patients was taken of the380 randomized into the trial. Their caserecords were examined for modifications indoses and for the reason for the changes. Thesample was chosen in a way which wouldcontribute proportionate numbers from eachof the participating clinics and would span thetime period during which the patients wereentered into the trial.Doses were modified in nine (16%) of thecontrol sample and in 36 (53%) of the treatedsample (table 13). In the placebo group thereasons for changing doses included thefollowing side effects: headache, nasal stuffi-ness, angina, skin rash, and multiple com-plaints. In the treated group 28 of the 36patients whose regimens were modified hadtheir doses reduced mainly because of hypo-tensive symptoms of weakness, lethargy, orfaintness associated with a fall of diastolicblood pressure to levels usually well below 90mm Hg. These symptoms abated followingappropriate reductions in doses. Side effectsother than hypotensive symptoms leading tochanges in doses included angina in threepatients, headache in three, nervousness inone, anid nasal stuffiness in one patient. Thus,the incidence of side effects other thanhypotension leading to changes in doses wasnot significaintly different in the control andtreated groups. Hypotensive symptoms prob-ably would have been far less if treatment hadbeeni initiated with a single antihypertensiveagent.Incidence of Subjective Side EffectsSubjective side effects in the same subsam-ple of 124 patients were evaluated from aphysician's interview checklist contained onthe clinic visit report forms. The reports forthe last two visits during the prerandomiza-tion trial period and all postrandomizationvisits were reviewed.Probably because of different value judg-ments used by the various clinic physiciansthere was considerable variation from onehospital to another regarding the frequencywith which specific complaints were reported.Since the same physician always saw aparticular patient it seemed valid to comparethe postrandomization against the prerandom-ization period for each patient. In anyparticular patient only those side effects werecounted which were complained of solelyduring the postrandomization period andwhich were not nioted prior to randomization The results indicate a surprising number ofpatients complaining of specific side effects inthe group of patients randomized on placebos.For example, of 52 patients in the controlgroup who did not complain of nightmaresprior to randomization, six reported this "sideeffect" at some time following randomization(table 14). For the specific complaints ofnightmares, arthritis, angina, and headache,the incidence actually was greater in thecontrol group of patients than in the treatedgroup. The incidence was nearly equal forboth control and treated groups for the sideeffects of depression, skin rash, impotence,and "other complaints." A possible explanationfor the high incidence of "side effects" is thatthe physician may have paid more attention tothe side effects interview following randomi-zation.The only side effects occurring with greaterfrequency in the treated group of patientswere lethargy or weakness, nasal stuffiness,ulcer symptoms, and first appearance of anycomplaint. The latter refers to the patientswho had no complaints of any kind prior torandomization but who reported one or moreside effects following randomization.It also seemed important to examine theincidence of side effects after appropriatemaintainence doses of the antihypertensivedrugs had been obtained. Therefore, datarelating to the reporting of specific side effectsomitting the first two postrandomization visits(third and subsequent visits only) also arepresented in table 14. For these latter visitsthe reporting of complaints of ulcer symp-toms, lethargy or weakness, and nasal stuffi-ness were insignificantly differeint in thecontrol and treated groups of patients. It isinteresting that during these visits the report-ing of angina and headache was considerablyhigher for the control as compared to thetreated group of patients.DiscussionThe incidence of morbid events in thepresent study was higher than o7ould beexpected in the general population of hyper-tensive patients. The reasons for this probablyinclude the following: (1) patients withhypertension on admission but whose diastolicblood pressure averaged below 90 mm Hgduring the fourth through sixth day ofhospitalization were excluded; (2) more thanhalf of the patients presented with cardiovas-cular or renal abnormalities; (3) in almost 30%of the patients hypertension was known to bepresenit for 10 years or longer; and (4) while only one fifth of the patients were above 60years of age such patients contributed half ofthe morbid events.The incidence of morbid events in thecontrol group was greater in patients withinitial levels of diastolic blood pressureaveraging 105-114 mm Hg than in those with90-104 mm Hg at entry. The effectiveness oftreatment was much greater for those with thehigher initial blood pressure levels. In thepresent communication it is shown thatpreexisting cardiovascular disease markedlyincreases the risk of developing events in thecontrol group and that the effectiveness oftreatment over the limited period of observa-tion was greater in those with such evidenceof prior disease.Termination of the trial was necessitated bythe clear-cut evidence of benefit in the treatedgroup as compared to the controls. Since thepatients without evidence of cardiovasculardisease were at reduced risk, follow-up was oftoo brief duration for the occurrence of manyevents or to obtain a statistically significantdifference between control and treated groupsin this subsample. However, the observed 50%effectiveness of treatment in the "no abnor-mality" group, which was not substantiallydifferent from the 64% effectiveness found inthe group with abnormalities, is consistentwith a protective effect of treatment.The results of the present trial justify moreinten.;ive efforts to identify and maintainunder adequate treatment patients with anysigns of cardiovascular damage or withdiastolic blood pressure averaging in excess of104 mm Hg. Evidence from surveys carriedout in representative population groups"6indicate that many of these patients are eitherunaware of their hypertension or are notreceiving adequate treatment.Additional evidence will be required, how-ever, to determine whether the benefits oftreatment outweigh its disadvantages in lowerrisk patients, such as in those with mildhypertension and no evidence of vasculardisease, particularly in women, and in patientswith labile hypertension. Toxic reactions mayoccur with any of the presently availableCirculation, Volume XLV, May 1972aintihypertensive agents, and side effect, par-ticularly biochemical changes associated withthiazides and related diuretics, are relativelycommon. Modifications of doses often arerequired because of other disturbing sideeffects. While the risk associated with thesevarious side effects appears small, it must beconsidered in relation to the benefit to beexpected in treating patients whose risk ofdeveloping complications due to hypertensionalso is relatively low. If patients are nlottreated, however, they should be followedperiodically to determine whether the hyper-tenision progresses to a more severe stage.Such follow-up appears to be particularlyimportant in younger patients. In the presentstudy, 15 of the 20 control patients whosediastolic blood pressures became severelyelevated were below 50 years of age; sevenihad an initial diastolic blood pressure below105 mm Hg.In the present sample of patienits, antihy-pertenisive treatment appeared to be effectivein reducing the complications associated withhypertension except for myocardial inifarctionanid sudden death. This result is not niecessar-ily iniconsistent with the statistical evidencethat elevated blood pressure is one of the "riskfactors" associated with an increased inci-denice of coronary heart disease. It is possiblethat a larger sample size or a longer period offollow-up might have revealed differences notapparent in the presenit study. Also, a greaterdegree of protection might have beeni affordedif treatment had beeni instituted at an earlierstage of hypertension. Therapeutic trials areneeded in a different population of hyperten-siv'e patients in order to resolve this question