Brian S. Alper, MD

Family physicians and general internists report an average of 6 questions for each half-day of office practice,^1-3 and 70% of these questions remain unanswered. The 2 factors that significantly predict whether a physician will attempt to answer a clinical question are the physician’s belief that a definitive answer exists and the urgency of the patient’s problem.⁴

Gorman and colleagues³ reported that medical librarians found clear answers for 46% of 60 randomly selected questions from family physicians; 51% would affect practice. The medical librarians searched for an average of 43 minutes per question. In a second study,⁵ medical librarians used MEDLINE and textbooks to answer 86 questions from family physicians. The MEDLINE searches took a mean of 27 minutes, and textbook searches took a mean of 6 minutes. Search results answered 54% of the clinical questions completely or nearly completely. Physicians estimated that the answers would have a “major” or “fairly major” impact on practice for 35% of their questions. MEDLINE searches provided answers to 43% of the questions, while textbook searches provided answers for an additional 11%.

Many physicians do not have the searching skills or access to the range of knowledge resources that librarians use. Even if they did, they do not take the time to conduct such searches during patient care. One study¹ found that physicians spent less than 2 minutes on average seeking an answer to a question. Thus, most clinical questions remain unanswered.

Electronic medical databases that provide answers directly (not just reference citations) may make it easier for clinicians to obtain answers at the point of care. We found no systematic evaluation of the capacity of such databases to answer clinical questions. We conducted this study to determine the extent to which current electronic medical databases can answer family physicians’ point-of-care clinical questions.

Methods

Database Selection

We solicited nominations for potentially suitable databases from multiple E-mail lists (including communities of family physicians [Family-L], medical informaticians [FAM-MED] and medical librarians [MEDLIB-L, MCMLA-L]) and through Web searches. A selection team consisting of 3 family physicians (J.S., D.W., B.E.) and a medical librarian (none of whom had financial relationships with any databases) determined whether the nominated databases met our inclusion criteria Table 1.

Clinical Questions

More than 1200 clinical questions had been previously collected from observations of family physicians during office practice.^1,5 These questions had been classified by typology (eg, Is test X indicated in situation Y?) and by topic (eg, dermatology).¹ We selected questions from these sources that were categorized among the most common typologies (8 of 68 typologies covering 50% of the questions) and the most common topics (7 of 62 topics covering 43% of the questions). These combinations of typologies and topics accounted for 272 (23%) of the 1204 questions.

If necessary, each question was translated by 2 physicians (B.A. and D.W. working together) to meet the following criteria: (1) clear enough to imagine an applicable clinical scenario, (2) answerable (ie, the question could theoretically be answered using clinical references without further patient data regardless of whether an answer was known to exist), (3) clinically relevant, and (4) true to the original question (ie, containing the information need and the modifying factors of the original question).

Each question was then independently proofread by at least 2 other physicians and translated again if necessary. Thirteen questions (5%) that did not meet these criteria after a second translation were dropped. Forty-seven questions (17%) that referred to information needs that could be adequately answered using the Physicians’ Desk Reference⁶ were dropped (eg, Are Paxil tablets scored?). The remaining 212 questions represented 8 typologies.¹ Two or 3 questions were randomly selected from each typology for a total of 20 questions Table 2.

Testing

Two family physicians with experience in computer searching (B.A., D.W.) independently searched for answers using each of the included databases. In the case of DynaMed, for which Dr Alper is the medical director, another family physician was substituted as a searcher, and Dr Alper had no input or control over the testing or arbitration process for answers from DynaMed. Testing took place in April and May 2000.

Searching was performed using computers with Pentium III processors with a 100 megabyte-per-second network connection to the Internet and server-mounted CD-ROMs.

Each searcher used the same 20 questions to evaluate each database. The order of evaluation of the databases was at the discretion of the searchers, but the testing of a database was completed before starting the testing of another database. Searchers became familiar with each database before testing it by using the 5 screening questions.

A maximum of 10 minutes was allowed per question. Each answer was rated as adequate or inadequate. An answer was considered adequate if it contained sufficient information to guide clinical practice. For example, for the question “How do I determine the cause of chronic pruritus?”, the answer from the University of Iowa Family Practice Handbook (www.vh.org/Providers/ClinRef/FPHandbook/Chapter13/01-13.html) was considered adequate, because it included clinically useful recommendations: History should include details about (1) any skin lesions preceding the pruritus; (2) history of weight loss, fatigue, fever, malaise; (3) any recent stress emotionally; and (4) recent medications and travel. Physical examination with emphasis on the skin and its appendages — xerosis, excoriation, lichenification, hydration. Laboratory tests as suggested by the PE, which may include CBC, ESR, fasting glucose, renal or liver function tests, hepatitis panel, thyroid tests, stool for parasites, CXR.

Sources that provided general recommendations without information that could specifically guide clinical practice were considered inadequate. For example: “The cause of generalized pruritus should be sought and corrected. If no skin disease is apparent, a systemic disorder or drug-related cause should be sought.” The searcher recorded the answer and the time it took to obtain it rounded to the nearest number of minutes (1-10).

Scoring and Arbitration

The 2 physician searchers judged the adequacy of the answers to each question for each database. If the searchers both found adequate answers, the result was accepted as adequate, and the average time required to find and interpret the answer was recorded. If neither searcher found an adequate answer, then the answer was deemed inadequate. If only one searcher found an adequate answer, the second searcher evaluated that answer. If the answer was acceptable to the second searcher, it was considered an adequate answer, and the time for the first searcher was recorded.

When searchers disagreed on the adequacy of identified answers, an arbitration panel consisting of 3 family physicians who were not affiliated with any of the databases met independently from the searchers to determine the adequacy of the answers by consensus.

Analysis

Our primary outcome was the proportion of questions adequately answered by each database. We calculated 95% confidence limits for the proportions of adequate answers.⁷ Means and medians were determined for the time to reach adequate answers for each database. We calculated the k statistic for the independent findings of the 2 searchers and for the results after the searchers reviewed each other’s searches.⁸ We combined the results of individual databases to determine the proportion of questions answered by all combinations of 2, 3, and 4 databases. We considered the question adequately answered if any of the individual databases adequately answered the question.

Results

Thirty-eight databases were nominated, and 24 did not meet our inclusion criteria Table W1.* Fourteen databases met the inclusion criteria Table 3 and were evaluated with the set of 20 questions (280 answer assessments) by 2 searchers. The Figure summarizes the process of evaluating the answers. The initial agreement between searchers was good k=0.69). Discussion between the searchers resolved 21 (52.5%) of the 40 discrepant answer assessments. These were due to inadequate searching or timing out (searching for 10 minutes) by one searcher, who agreed with the adequacy of the answer found by the other searcher. The agreement between searchers at this stage was excellent (k= 0.94).

The remaining 19 discrepant assessments (for which the searchers had different opinions regarding the adequacy of the answers identified) were referred to the arbitration panel for determination of the final results. Ten of these were deemed adequate.

Results for individual databases in rank order of proportion of questions answered followed by average time to identify adequate answers are reported in Table 3. The combination of STAT!Ref and MDConsult could answer 85% of our set of 20 questions. Four combinations of 2 databases (STAT!Ref and either MAXX, MDChoice.com, Primary Care Guidelines, or Medscape) could answer 80% of our questions. Two combinations of 3 databases (STAT!Ref, MDConsult, and either DynaMed or MAXX) could answer 90% of our questions. Combinations of 4 databases answered the most sample questions (95%, 19/20). These combinations consisted of STAT!Ref, DynaMed, MAXX, and either MDConsult or American Family Physician.

We also evaluated combinations of databases that were available at no cost. The combination of the 2 no-cost databases that answered the largest proportion of questions (75%) was DynaMed and American Family Physician. The greatest proportion of clinical questions that could be answered using the freely available sources was 80%, and this required the use of 3 databases (DynaMed, MDChoice.com, and American Family Physician).

Discussion

Our study suggests that individual databases can answer a considerable proportion of family physicians’ clinical questions. Combinations of currently available databases can answer 75% or more. The searches in this study were based on the combination of efforts of 2 experienced physician searchers. These results may not be replicable in the practice setting but do provide an objective best-case scenario assessment of the content of these databases.

The time required to obtain answers, while much less than searching for original articles, is still longer than the 2-minute average time spent by family physicians in the study by Ely and colleagues.¹ Our time estimates are not precise, as time was not the primary focus of our study. Time was only recorded in 1-minute intervals, so searches that took 10 seconds were recorded as 1 minute. Even so, the existence of median times to obtain adequate answers greater than 2 minutes suggests that these databases may require more time than most physicians will take to pursue answers during patient care.

This is the first study to systematically evaluate how many questions can be answered by electronic medical databases. The strengths of this study include the use of a standard set of common questions asked by family physicians, testing by 2 experienced family physician searchers, and a systematic replicable approach to the evaluation. The only similar study we identified was one in which Graber and coworkers⁹ used 10 clinical questions and tested a commercial site, 2 medical meta-lists, 4 general search engines, and 9 medicine-specific search engines to determine the efficiency of answering clinical questions on the Web. Different approaches answered from 0 to 6 of the 10 questions, but that study looked primarily at sites that were not generally designed for use in clinical practice.

Limitations

Our study was limited by the relatively small number of questions, causing wide confidence intervals. Some answers were present in the databases but not found despite the use of 2 searchers. For example, a database manager identified 2 answers that were not found but would have been considered adequate.

We accepted answers as adequate if, in our judgment, they offered a practical course of action. We did not attempt to determine whether the individual asking the question believed that the answer was adequate nor did we attempt to validate the accuracy or currency of answers using independent standards. Many of the answers were based on sources that were several years old, and few were based on explicit evidence-based criteria. Although we determined the adequacy of answers for clinical practice through formal mechanisms, an in vivo study in which the clinicians asking the questions determined the adequacy of their findings during patient care activities would provide a more accurate assessment.

Our study presents a static evaluation of a dynamic field. Over time, answers may be lost because of lack of maintenance of resource links or may be gained by addition of new materials. Our use of questions gathered several years ago may not accurately reflect the ability of databases to answer current questions, which may be more likely to reflect new tests and treatments.

Many of the databases were designed for purposes other than meeting clinical information needs at the point of care. Performance in this study does not reflect the capacity of these databases to address their stated purposes. For example, the Translating Research Into Practice (TRIP) database is an excellent resource for searches of a large collection of evidence-based resources. These resources are generally limited to summaries of studies with the highest methodologic quality. The TRIP database did not perform well in our study partly because most of our test questions (consistent with questions in clinical practice) cannot currently be answered using studies of the highest methodologic quality. Another example is Medical Matrix, which provides a search engine and annotated summaries for exploring the entire medical Internet and not just clinical reference information.

We did not study the costs involved in using the databases we evaluated, and these costs may have changed since our study was conducted. Most of the databases we included were free to use at the time of the study and at the time of this report. The 3 collections of textbooks required access fees. STAT!Ref, which scored the highest in our study, did so because we used the complete collection available to us through our institutional library. This collection would cost an individual $2189 annually at the time of our study. A starter library was available for $199 annually and would only answer 40% of the questions.

Context

Family physicians and other primary care providers treat patients who have a wide variety of syndromes and symptoms. Because of the scope and breadth of primary care, it is nearly impossible for a clinician to keep up with rapidly changing medical information.¹⁰

Connelly and colleagues¹¹ surveyed 126 family physicians and found they used the Physicians’ Desk Reference and colleagues much more often than Index Medicus or computer-based bibliographic retrieval systems. Research literature was used infrequently and rated among the lowest in terms of credibility, availability, searchability, understandability, and applicability. Physicians preferred sources that had low cost and were relevant to specific patient problems over sources that had higher quality.

Conclusions

Current databases can answer a considerable proportion of clinical questions but have not reached their potential for efficiency. It is our hope that as electronic medical databases mature, they will be able to bridge this gap and bring the research literature to the point of care in useful and practical ways. This study provides a snapshot of how far we have come and how far we need to go to meet these needs.

Acknowledgments

Funding for our study was provided by a grant from the American Academy of Family Physicians to support the Center for Family Medicine Science and from 2 Bureau of Health Professions Awards (DHHS 1-D14-HP-00029-01, DHHS 5 T32 HP10038) from the Health Resources and Services Administration to the Department of Family and Community Medicine at University of Missouri-Columbia. The authors would like to acknowledge Erik Lindbloom, MD, MSPH, for assisting with the database testing as a substitute searcher for B.A.; E. Diane Johnson, MLS, for assisting with the selection of databases for study inclusion; Robert Phillips, Jr., MD, MSPH, for arbitration of questions and answers for which the searchers did not reach agreement along with B.E. and J.S.; David Cravens, Erik Lindbloom, Kevin Kane, Jim Brillhart, and Mark Ebell for proofreading the questions for clarity, answerability, and clinical relevance; John Ely and Lee Chambliss for providing clinical questions from their observations; Mark Ebell, John Ely, Erik Lindbloom, Jerry Osheroff, Lee Chambliss, David Mehr, Robin Kruse, John Smucny, and many others for constructive criticism in the design of this study; and Steve Zweig for editorial review.

Family physicians and general internists report an average of 6 questions for each half-day of office practice,^1-3 and 70% of these questions remain unanswered. The 2 factors that significantly predict whether a physician will attempt to answer a clinical question are the physician’s belief that a definitive answer exists and the urgency of the patient’s problem.⁴

Gorman and colleagues³ reported that medical librarians found clear answers for 46% of 60 randomly selected questions from family physicians; 51% would affect practice. The medical librarians searched for an average of 43 minutes per question. In a second study,⁵ medical librarians used MEDLINE and textbooks to answer 86 questions from family physicians. The MEDLINE searches took a mean of 27 minutes, and textbook searches took a mean of 6 minutes. Search results answered 54% of the clinical questions completely or nearly completely. Physicians estimated that the answers would have a “major” or “fairly major” impact on practice for 35% of their questions. MEDLINE searches provided answers to 43% of the questions, while textbook searches provided answers for an additional 11%.

Many physicians do not have the searching skills or access to the range of knowledge resources that librarians use. Even if they did, they do not take the time to conduct such searches during patient care. One study¹ found that physicians spent less than 2 minutes on average seeking an answer to a question. Thus, most clinical questions remain unanswered.

Electronic medical databases that provide answers directly (not just reference citations) may make it easier for clinicians to obtain answers at the point of care. We found no systematic evaluation of the capacity of such databases to answer clinical questions. We conducted this study to determine the extent to which current electronic medical databases can answer family physicians’ point-of-care clinical questions.

Methods

Database Selection

We solicited nominations for potentially suitable databases from multiple E-mail lists (including communities of family physicians [Family-L], medical informaticians [FAM-MED] and medical librarians [MEDLIB-L, MCMLA-L]) and through Web searches. A selection team consisting of 3 family physicians (J.S., D.W., B.E.) and a medical librarian (none of whom had financial relationships with any databases) determined whether the nominated databases met our inclusion criteria Table 1.

Clinical Questions

More than 1200 clinical questions had been previously collected from observations of family physicians during office practice.^1,5 These questions had been classified by typology (eg, Is test X indicated in situation Y?) and by topic (eg, dermatology).¹ We selected questions from these sources that were categorized among the most common typologies (8 of 68 typologies covering 50% of the questions) and the most common topics (7 of 62 topics covering 43% of the questions). These combinations of typologies and topics accounted for 272 (23%) of the 1204 questions.

If necessary, each question was translated by 2 physicians (B.A. and D.W. working together) to meet the following criteria: (1) clear enough to imagine an applicable clinical scenario, (2) answerable (ie, the question could theoretically be answered using clinical references without further patient data regardless of whether an answer was known to exist), (3) clinically relevant, and (4) true to the original question (ie, containing the information need and the modifying factors of the original question).

Each question was then independently proofread by at least 2 other physicians and translated again if necessary. Thirteen questions (5%) that did not meet these criteria after a second translation were dropped. Forty-seven questions (17%) that referred to information needs that could be adequately answered using the Physicians’ Desk Reference⁶ were dropped (eg, Are Paxil tablets scored?). The remaining 212 questions represented 8 typologies.¹ Two or 3 questions were randomly selected from each typology for a total of 20 questions Table 2.

Testing

Two family physicians with experience in computer searching (B.A., D.W.) independently searched for answers using each of the included databases. In the case of DynaMed, for which Dr Alper is the medical director, another family physician was substituted as a searcher, and Dr Alper had no input or control over the testing or arbitration process for answers from DynaMed. Testing took place in April and May 2000.

Searching was performed using computers with Pentium III processors with a 100 megabyte-per-second network connection to the Internet and server-mounted CD-ROMs.

Each searcher used the same 20 questions to evaluate each database. The order of evaluation of the databases was at the discretion of the searchers, but the testing of a database was completed before starting the testing of another database. Searchers became familiar with each database before testing it by using the 5 screening questions.

A maximum of 10 minutes was allowed per question. Each answer was rated as adequate or inadequate. An answer was considered adequate if it contained sufficient information to guide clinical practice. For example, for the question “How do I determine the cause of chronic pruritus?”, the answer from the University of Iowa Family Practice Handbook (www.vh.org/Providers/ClinRef/FPHandbook/Chapter13/01-13.html) was considered adequate, because it included clinically useful recommendations: History should include details about (1) any skin lesions preceding the pruritus; (2) history of weight loss, fatigue, fever, malaise; (3) any recent stress emotionally; and (4) recent medications and travel. Physical examination with emphasis on the skin and its appendages — xerosis, excoriation, lichenification, hydration. Laboratory tests as suggested by the PE, which may include CBC, ESR, fasting glucose, renal or liver function tests, hepatitis panel, thyroid tests, stool for parasites, CXR.

Sources that provided general recommendations without information that could specifically guide clinical practice were considered inadequate. For example: “The cause of generalized pruritus should be sought and corrected. If no skin disease is apparent, a systemic disorder or drug-related cause should be sought.” The searcher recorded the answer and the time it took to obtain it rounded to the nearest number of minutes (1-10).

Scoring and Arbitration

The 2 physician searchers judged the adequacy of the answers to each question for each database. If the searchers both found adequate answers, the result was accepted as adequate, and the average time required to find and interpret the answer was recorded. If neither searcher found an adequate answer, then the answer was deemed inadequate. If only one searcher found an adequate answer, the second searcher evaluated that answer. If the answer was acceptable to the second searcher, it was considered an adequate answer, and the time for the first searcher was recorded.

When searchers disagreed on the adequacy of identified answers, an arbitration panel consisting of 3 family physicians who were not affiliated with any of the databases met independently from the searchers to determine the adequacy of the answers by consensus.

Analysis

Our primary outcome was the proportion of questions adequately answered by each database. We calculated 95% confidence limits for the proportions of adequate answers.⁷ Means and medians were determined for the time to reach adequate answers for each database. We calculated the k statistic for the independent findings of the 2 searchers and for the results after the searchers reviewed each other’s searches.⁸ We combined the results of individual databases to determine the proportion of questions answered by all combinations of 2, 3, and 4 databases. We considered the question adequately answered if any of the individual databases adequately answered the question.

Results

Thirty-eight databases were nominated, and 24 did not meet our inclusion criteria Table W1.* Fourteen databases met the inclusion criteria Table 3 and were evaluated with the set of 20 questions (280 answer assessments) by 2 searchers. The Figure summarizes the process of evaluating the answers. The initial agreement between searchers was good k=0.69). Discussion between the searchers resolved 21 (52.5%) of the 40 discrepant answer assessments. These were due to inadequate searching or timing out (searching for 10 minutes) by one searcher, who agreed with the adequacy of the answer found by the other searcher. The agreement between searchers at this stage was excellent (k= 0.94).

The remaining 19 discrepant assessments (for which the searchers had different opinions regarding the adequacy of the answers identified) were referred to the arbitration panel for determination of the final results. Ten of these were deemed adequate.

Results for individual databases in rank order of proportion of questions answered followed by average time to identify adequate answers are reported in Table 3. The combination of STAT!Ref and MDConsult could answer 85% of our set of 20 questions. Four combinations of 2 databases (STAT!Ref and either MAXX, MDChoice.com, Primary Care Guidelines, or Medscape) could answer 80% of our questions. Two combinations of 3 databases (STAT!Ref, MDConsult, and either DynaMed or MAXX) could answer 90% of our questions. Combinations of 4 databases answered the most sample questions (95%, 19/20). These combinations consisted of STAT!Ref, DynaMed, MAXX, and either MDConsult or American Family Physician.

We also evaluated combinations of databases that were available at no cost. The combination of the 2 no-cost databases that answered the largest proportion of questions (75%) was DynaMed and American Family Physician. The greatest proportion of clinical questions that could be answered using the freely available sources was 80%, and this required the use of 3 databases (DynaMed, MDChoice.com, and American Family Physician).

Discussion

Our study suggests that individual databases can answer a considerable proportion of family physicians’ clinical questions. Combinations of currently available databases can answer 75% or more. The searches in this study were based on the combination of efforts of 2 experienced physician searchers. These results may not be replicable in the practice setting but do provide an objective best-case scenario assessment of the content of these databases.

The time required to obtain answers, while much less than searching for original articles, is still longer than the 2-minute average time spent by family physicians in the study by Ely and colleagues.¹ Our time estimates are not precise, as time was not the primary focus of our study. Time was only recorded in 1-minute intervals, so searches that took 10 seconds were recorded as 1 minute. Even so, the existence of median times to obtain adequate answers greater than 2 minutes suggests that these databases may require more time than most physicians will take to pursue answers during patient care.

This is the first study to systematically evaluate how many questions can be answered by electronic medical databases. The strengths of this study include the use of a standard set of common questions asked by family physicians, testing by 2 experienced family physician searchers, and a systematic replicable approach to the evaluation. The only similar study we identified was one in which Graber and coworkers⁹ used 10 clinical questions and tested a commercial site, 2 medical meta-lists, 4 general search engines, and 9 medicine-specific search engines to determine the efficiency of answering clinical questions on the Web. Different approaches answered from 0 to 6 of the 10 questions, but that study looked primarily at sites that were not generally designed for use in clinical practice.

Limitations

Our study was limited by the relatively small number of questions, causing wide confidence intervals. Some answers were present in the databases but not found despite the use of 2 searchers. For example, a database manager identified 2 answers that were not found but would have been considered adequate.

We accepted answers as adequate if, in our judgment, they offered a practical course of action. We did not attempt to determine whether the individual asking the question believed that the answer was adequate nor did we attempt to validate the accuracy or currency of answers using independent standards. Many of the answers were based on sources that were several years old, and few were based on explicit evidence-based criteria. Although we determined the adequacy of answers for clinical practice through formal mechanisms, an in vivo study in which the clinicians asking the questions determined the adequacy of their findings during patient care activities would provide a more accurate assessment.

Our study presents a static evaluation of a dynamic field. Over time, answers may be lost because of lack of maintenance of resource links or may be gained by addition of new materials. Our use of questions gathered several years ago may not accurately reflect the ability of databases to answer current questions, which may be more likely to reflect new tests and treatments.

Many of the databases were designed for purposes other than meeting clinical information needs at the point of care. Performance in this study does not reflect the capacity of these databases to address their stated purposes. For example, the Translating Research Into Practice (TRIP) database is an excellent resource for searches of a large collection of evidence-based resources. These resources are generally limited to summaries of studies with the highest methodologic quality. The TRIP database did not perform well in our study partly because most of our test questions (consistent with questions in clinical practice) cannot currently be answered using studies of the highest methodologic quality. Another example is Medical Matrix, which provides a search engine and annotated summaries for exploring the entire medical Internet and not just clinical reference information.

We did not study the costs involved in using the databases we evaluated, and these costs may have changed since our study was conducted. Most of the databases we included were free to use at the time of the study and at the time of this report. The 3 collections of textbooks required access fees. STAT!Ref, which scored the highest in our study, did so because we used the complete collection available to us through our institutional library. This collection would cost an individual $2189 annually at the time of our study. A starter library was available for $199 annually and would only answer 40% of the questions.

Context

Family physicians and other primary care providers treat patients who have a wide variety of syndromes and symptoms. Because of the scope and breadth of primary care, it is nearly impossible for a clinician to keep up with rapidly changing medical information.¹⁰

Connelly and colleagues¹¹ surveyed 126 family physicians and found they used the Physicians’ Desk Reference and colleagues much more often than Index Medicus or computer-based bibliographic retrieval systems. Research literature was used infrequently and rated among the lowest in terms of credibility, availability, searchability, understandability, and applicability. Physicians preferred sources that had low cost and were relevant to specific patient problems over sources that had higher quality.

Conclusions

Current databases can answer a considerable proportion of clinical questions but have not reached their potential for efficiency. It is our hope that as electronic medical databases mature, they will be able to bridge this gap and bring the research literature to the point of care in useful and practical ways. This study provides a snapshot of how far we have come and how far we need to go to meet these needs.

Acknowledgments

Funding for our study was provided by a grant from the American Academy of Family Physicians to support the Center for Family Medicine Science and from 2 Bureau of Health Professions Awards (DHHS 1-D14-HP-00029-01, DHHS 5 T32 HP10038) from the Health Resources and Services Administration to the Department of Family and Community Medicine at University of Missouri-Columbia. The authors would like to acknowledge Erik Lindbloom, MD, MSPH, for assisting with the database testing as a substitute searcher for B.A.; E. Diane Johnson, MLS, for assisting with the selection of databases for study inclusion; Robert Phillips, Jr., MD, MSPH, for arbitration of questions and answers for which the searchers did not reach agreement along with B.E. and J.S.; David Cravens, Erik Lindbloom, Kevin Kane, Jim Brillhart, and Mark Ebell for proofreading the questions for clarity, answerability, and clinical relevance; John Ely and Lee Chambliss for providing clinical questions from their observations; Mark Ebell, John Ely, Erik Lindbloom, Jerry Osheroff, Lee Chambliss, David Mehr, Robin Kruse, John Smucny, and many others for constructive criticism in the design of this study; and Steve Zweig for editorial review.

Family physicians and general internists report an average of 6 questions for each half-day of office practice,^1-3 and 70% of these questions remain unanswered. The 2 factors that significantly predict whether a physician will attempt to answer a clinical question are the physician’s belief that a definitive answer exists and the urgency of the patient’s problem.⁴

Gorman and colleagues³ reported that medical librarians found clear answers for 46% of 60 randomly selected questions from family physicians; 51% would affect practice. The medical librarians searched for an average of 43 minutes per question. In a second study,⁵ medical librarians used MEDLINE and textbooks to answer 86 questions from family physicians. The MEDLINE searches took a mean of 27 minutes, and textbook searches took a mean of 6 minutes. Search results answered 54% of the clinical questions completely or nearly completely. Physicians estimated that the answers would have a “major” or “fairly major” impact on practice for 35% of their questions. MEDLINE searches provided answers to 43% of the questions, while textbook searches provided answers for an additional 11%.

Many physicians do not have the searching skills or access to the range of knowledge resources that librarians use. Even if they did, they do not take the time to conduct such searches during patient care. One study¹ found that physicians spent less than 2 minutes on average seeking an answer to a question. Thus, most clinical questions remain unanswered.

Electronic medical databases that provide answers directly (not just reference citations) may make it easier for clinicians to obtain answers at the point of care. We found no systematic evaluation of the capacity of such databases to answer clinical questions. We conducted this study to determine the extent to which current electronic medical databases can answer family physicians’ point-of-care clinical questions.

Methods

Database Selection

We solicited nominations for potentially suitable databases from multiple E-mail lists (including communities of family physicians [Family-L], medical informaticians [FAM-MED] and medical librarians [MEDLIB-L, MCMLA-L]) and through Web searches. A selection team consisting of 3 family physicians (J.S., D.W., B.E.) and a medical librarian (none of whom had financial relationships with any databases) determined whether the nominated databases met our inclusion criteria Table 1.

Clinical Questions

More than 1200 clinical questions had been previously collected from observations of family physicians during office practice.^1,5 These questions had been classified by typology (eg, Is test X indicated in situation Y?) and by topic (eg, dermatology).¹ We selected questions from these sources that were categorized among the most common typologies (8 of 68 typologies covering 50% of the questions) and the most common topics (7 of 62 topics covering 43% of the questions). These combinations of typologies and topics accounted for 272 (23%) of the 1204 questions.

If necessary, each question was translated by 2 physicians (B.A. and D.W. working together) to meet the following criteria: (1) clear enough to imagine an applicable clinical scenario, (2) answerable (ie, the question could theoretically be answered using clinical references without further patient data regardless of whether an answer was known to exist), (3) clinically relevant, and (4) true to the original question (ie, containing the information need and the modifying factors of the original question).

Each question was then independently proofread by at least 2 other physicians and translated again if necessary. Thirteen questions (5%) that did not meet these criteria after a second translation were dropped. Forty-seven questions (17%) that referred to information needs that could be adequately answered using the Physicians’ Desk Reference⁶ were dropped (eg, Are Paxil tablets scored?). The remaining 212 questions represented 8 typologies.¹ Two or 3 questions were randomly selected from each typology for a total of 20 questions Table 2.

Testing

Two family physicians with experience in computer searching (B.A., D.W.) independently searched for answers using each of the included databases. In the case of DynaMed, for which Dr Alper is the medical director, another family physician was substituted as a searcher, and Dr Alper had no input or control over the testing or arbitration process for answers from DynaMed. Testing took place in April and May 2000.

Searching was performed using computers with Pentium III processors with a 100 megabyte-per-second network connection to the Internet and server-mounted CD-ROMs.

Each searcher used the same 20 questions to evaluate each database. The order of evaluation of the databases was at the discretion of the searchers, but the testing of a database was completed before starting the testing of another database. Searchers became familiar with each database before testing it by using the 5 screening questions.

A maximum of 10 minutes was allowed per question. Each answer was rated as adequate or inadequate. An answer was considered adequate if it contained sufficient information to guide clinical practice. For example, for the question “How do I determine the cause of chronic pruritus?”, the answer from the University of Iowa Family Practice Handbook (www.vh.org/Providers/ClinRef/FPHandbook/Chapter13/01-13.html) was considered adequate, because it included clinically useful recommendations: History should include details about (1) any skin lesions preceding the pruritus; (2) history of weight loss, fatigue, fever, malaise; (3) any recent stress emotionally; and (4) recent medications and travel. Physical examination with emphasis on the skin and its appendages — xerosis, excoriation, lichenification, hydration. Laboratory tests as suggested by the PE, which may include CBC, ESR, fasting glucose, renal or liver function tests, hepatitis panel, thyroid tests, stool for parasites, CXR.

Sources that provided general recommendations without information that could specifically guide clinical practice were considered inadequate. For example: “The cause of generalized pruritus should be sought and corrected. If no skin disease is apparent, a systemic disorder or drug-related cause should be sought.” The searcher recorded the answer and the time it took to obtain it rounded to the nearest number of minutes (1-10).

Scoring and Arbitration

The 2 physician searchers judged the adequacy of the answers to each question for each database. If the searchers both found adequate answers, the result was accepted as adequate, and the average time required to find and interpret the answer was recorded. If neither searcher found an adequate answer, then the answer was deemed inadequate. If only one searcher found an adequate answer, the second searcher evaluated that answer. If the answer was acceptable to the second searcher, it was considered an adequate answer, and the time for the first searcher was recorded.

When searchers disagreed on the adequacy of identified answers, an arbitration panel consisting of 3 family physicians who were not affiliated with any of the databases met independently from the searchers to determine the adequacy of the answers by consensus.

Analysis

Our primary outcome was the proportion of questions adequately answered by each database. We calculated 95% confidence limits for the proportions of adequate answers.⁷ Means and medians were determined for the time to reach adequate answers for each database. We calculated the k statistic for the independent findings of the 2 searchers and for the results after the searchers reviewed each other’s searches.⁸ We combined the results of individual databases to determine the proportion of questions answered by all combinations of 2, 3, and 4 databases. We considered the question adequately answered if any of the individual databases adequately answered the question.

Results

Thirty-eight databases were nominated, and 24 did not meet our inclusion criteria Table W1.* Fourteen databases met the inclusion criteria Table 3 and were evaluated with the set of 20 questions (280 answer assessments) by 2 searchers. The Figure summarizes the process of evaluating the answers. The initial agreement between searchers was good k=0.69). Discussion between the searchers resolved 21 (52.5%) of the 40 discrepant answer assessments. These were due to inadequate searching or timing out (searching for 10 minutes) by one searcher, who agreed with the adequacy of the answer found by the other searcher. The agreement between searchers at this stage was excellent (k= 0.94).

The remaining 19 discrepant assessments (for which the searchers had different opinions regarding the adequacy of the answers identified) were referred to the arbitration panel for determination of the final results. Ten of these were deemed adequate.

Results for individual databases in rank order of proportion of questions answered followed by average time to identify adequate answers are reported in Table 3. The combination of STAT!Ref and MDConsult could answer 85% of our set of 20 questions. Four combinations of 2 databases (STAT!Ref and either MAXX, MDChoice.com, Primary Care Guidelines, or Medscape) could answer 80% of our questions. Two combinations of 3 databases (STAT!Ref, MDConsult, and either DynaMed or MAXX) could answer 90% of our questions. Combinations of 4 databases answered the most sample questions (95%, 19/20). These combinations consisted of STAT!Ref, DynaMed, MAXX, and either MDConsult or American Family Physician.

We also evaluated combinations of databases that were available at no cost. The combination of the 2 no-cost databases that answered the largest proportion of questions (75%) was DynaMed and American Family Physician. The greatest proportion of clinical questions that could be answered using the freely available sources was 80%, and this required the use of 3 databases (DynaMed, MDChoice.com, and American Family Physician).

Discussion

Our study suggests that individual databases can answer a considerable proportion of family physicians’ clinical questions. Combinations of currently available databases can answer 75% or more. The searches in this study were based on the combination of efforts of 2 experienced physician searchers. These results may not be replicable in the practice setting but do provide an objective best-case scenario assessment of the content of these databases.

The time required to obtain answers, while much less than searching for original articles, is still longer than the 2-minute average time spent by family physicians in the study by Ely and colleagues.¹ Our time estimates are not precise, as time was not the primary focus of our study. Time was only recorded in 1-minute intervals, so searches that took 10 seconds were recorded as 1 minute. Even so, the existence of median times to obtain adequate answers greater than 2 minutes suggests that these databases may require more time than most physicians will take to pursue answers during patient care.

This is the first study to systematically evaluate how many questions can be answered by electronic medical databases. The strengths of this study include the use of a standard set of common questions asked by family physicians, testing by 2 experienced family physician searchers, and a systematic replicable approach to the evaluation. The only similar study we identified was one in which Graber and coworkers⁹ used 10 clinical questions and tested a commercial site, 2 medical meta-lists, 4 general search engines, and 9 medicine-specific search engines to determine the efficiency of answering clinical questions on the Web. Different approaches answered from 0 to 6 of the 10 questions, but that study looked primarily at sites that were not generally designed for use in clinical practice.

Limitations

Our study was limited by the relatively small number of questions, causing wide confidence intervals. Some answers were present in the databases but not found despite the use of 2 searchers. For example, a database manager identified 2 answers that were not found but would have been considered adequate.

We accepted answers as adequate if, in our judgment, they offered a practical course of action. We did not attempt to determine whether the individual asking the question believed that the answer was adequate nor did we attempt to validate the accuracy or currency of answers using independent standards. Many of the answers were based on sources that were several years old, and few were based on explicit evidence-based criteria. Although we determined the adequacy of answers for clinical practice through formal mechanisms, an in vivo study in which the clinicians asking the questions determined the adequacy of their findings during patient care activities would provide a more accurate assessment.

Our study presents a static evaluation of a dynamic field. Over time, answers may be lost because of lack of maintenance of resource links or may be gained by addition of new materials. Our use of questions gathered several years ago may not accurately reflect the ability of databases to answer current questions, which may be more likely to reflect new tests and treatments.

Many of the databases were designed for purposes other than meeting clinical information needs at the point of care. Performance in this study does not reflect the capacity of these databases to address their stated purposes. For example, the Translating Research Into Practice (TRIP) database is an excellent resource for searches of a large collection of evidence-based resources. These resources are generally limited to summaries of studies with the highest methodologic quality. The TRIP database did not perform well in our study partly because most of our test questions (consistent with questions in clinical practice) cannot currently be answered using studies of the highest methodologic quality. Another example is Medical Matrix, which provides a search engine and annotated summaries for exploring the entire medical Internet and not just clinical reference information.

We did not study the costs involved in using the databases we evaluated, and these costs may have changed since our study was conducted. Most of the databases we included were free to use at the time of the study and at the time of this report. The 3 collections of textbooks required access fees. STAT!Ref, which scored the highest in our study, did so because we used the complete collection available to us through our institutional library. This collection would cost an individual $2189 annually at the time of our study. A starter library was available for $199 annually and would only answer 40% of the questions.

Context

Family physicians and other primary care providers treat patients who have a wide variety of syndromes and symptoms. Because of the scope and breadth of primary care, it is nearly impossible for a clinician to keep up with rapidly changing medical information.¹⁰

Connelly and colleagues¹¹ surveyed 126 family physicians and found they used the Physicians’ Desk Reference and colleagues much more often than Index Medicus or computer-based bibliographic retrieval systems. Research literature was used infrequently and rated among the lowest in terms of credibility, availability, searchability, understandability, and applicability. Physicians preferred sources that had low cost and were relevant to specific patient problems over sources that had higher quality.

Conclusions

Current databases can answer a considerable proportion of clinical questions but have not reached their potential for efficiency. It is our hope that as electronic medical databases mature, they will be able to bridge this gap and bring the research literature to the point of care in useful and practical ways. This study provides a snapshot of how far we have come and how far we need to go to meet these needs.

Acknowledgments

Funding for our study was provided by a grant from the American Academy of Family Physicians to support the Center for Family Medicine Science and from 2 Bureau of Health Professions Awards (DHHS 1-D14-HP-00029-01, DHHS 5 T32 HP10038) from the Health Resources and Services Administration to the Department of Family and Community Medicine at University of Missouri-Columbia. The authors would like to acknowledge Erik Lindbloom, MD, MSPH, for assisting with the database testing as a substitute searcher for B.A.; E. Diane Johnson, MLS, for assisting with the selection of databases for study inclusion; Robert Phillips, Jr., MD, MSPH, for arbitration of questions and answers for which the searchers did not reach agreement along with B.E. and J.S.; David Cravens, Erik Lindbloom, Kevin Kane, Jim Brillhart, and Mark Ebell for proofreading the questions for clarity, answerability, and clinical relevance; John Ely and Lee Chambliss for providing clinical questions from their observations; Mark Ebell, John Ely, Erik Lindbloom, Jerry Osheroff, Lee Chambliss, David Mehr, Robin Kruse, John Smucny, and many others for constructive criticism in the design of this study; and Steve Zweig for editorial review.

Standard obstetrical practice has included screening for anemia and the provision of iron supplements to anemic patients. This approach has been based on assumptions about anemia and iron deficiency that are not supported by the literature.

Anemia in pregnancy has been reported to be associated with preterm delivery.^1,2 However, this may not take into consideration the lower hemoglobin values normally present during the second trimester. The potentially spurious association between anemia and preterm delivery could be explained by the lower hemoglobin values that are expected during the second trimester. When measured at the time of preterm delivery, these lower hemoglobin values, which are normal in the second trimester (but not in the third trimester or in a nonpregnant woman), are often believed to have led to the preterm delivery.

Normal hemogloblin values from nonpregnant women cannot be assumed to apply to those who are pregnant. Average hemoglobin levels decrease to 11.6 g per dL at 20 to 24 weeks’ gestation, with the fifth percentile at 10.5 g per dL (hematocrit=32%). Anemia in pregnancy has been defined by criteria from the Centers for Disease Control and Prevention (CDC) as a hemoglobin level of less than 11 g per dL during the first and third trimesters and less than 10.5 g per dL during the second trimester.³

When anemia is present in pregnancy it cannot be assumed to be the result of iron deficiency, even though this type of anemia has been previously reported as the most common cause.⁴ The Camden study⁵ of 826 pregnant women showed preterm delivery and low birth weight associated with iron-deficiency anemia. Only 27.9% of the pregnant women had anemia, however, and only 12.5% of the patients with anemia had an iron deficiency. Thus, only 3.5% of the entire cohort had iron-deficiency anemia.⁶

Iron deficiency in pregnancy has been defined by the National Academy of Sciences panel on nutrition and pregnancy⁷ as ferritin levels lower than 12 ng per mL. A systematic overview⁸ of 55 studies relevant to laboratory tests for diagnosis of iron-deficiency anemia in variable patient populations found serum ferritin radioimmunoassay to be the most powerful test. Ferritin levels are considered the gold standard for the diagnosis of iron-deficiency anemia in pregnancy.⁹

We report a descriptive study of the use of ferritin levels to determine the need for iron supplementation among pregnant women with anemia.

Methods

Lebanon Family Health Services is a nonprofit federally subsidized community agency providing prenatal care and women’s health services to a diverse population without restriction on the basis of financial status. The prenatal patients cared for in this practice included women aged 15 to 40 years (23.6% were younger than 19 years; 70.8%, 19 to 30 years; and 5.6%, older than 30 years) of whom 65.3% were white, 29.2% Hispanic, 4.2% African American, and 1.4% Asian.

We evaluated all patients entering into prenatal care at Lebanon Family Health Services from April 1, 1997, through December 31, 1998, using prospective data collection and retrospective record review. Prenatal vitamins (including elemental iron 30 to 60 mg/day) were prescribed to all patients. Complete blood count was tested at the initial evaluation as part of a comprehensive screening. For patients who entered prenatal care at earlier than 28 weeks’ gestation complete blood count was checked again when they had reached that point.

The CDC criteria for defining anemia are hemoglobin levels less than 11 g per dL during the first and third trimesters, and less than 10.5 g per dL during the second trimester. A hemoglobin level of less than 11 g per dL at any time during the pregnancy was used as the cutoff point for anemia in this clinical practice, in consideration of patients with uncertain or inaccurate pregnancy dating.

Patients with anemia underwent blood testing for serum ferritin level, generally 1 week after a complete blood cell count. At the time, ferritin levels were determined using different test tubes than those used for other prenatal testing, thus resulting in a delay in obtaining blood for ferritin testing. If serum ferritin was 12 ng per dL or lower, iron-deficiency anemia was diagnosed and ferrous sulfate was prescribed for the remainder of the pregnancy and the postpartum period. If serum ferritin was greater than 12 ng per dL iron deficiency was excluded, anemia was generally considered to be pregnancy related, and further evaluation and treatment was at the discretion of the treating clinician.

Variables recorded for all patients with anemia were: estimated date of delivery, last menstrual period, date of any testing for complete blood count or ferritin levels, hemoglobin value, hematocrit value, red blood cell count (RBC), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), red cell distribution width (RDW), any notation regarding microcytosis, and serum ferritin levels.

Results

A total of 182 patients were consecutively entered into prenatal care during the study period. Hemoglobin data were not obtained from 9 patients—4 transferred to another practice for care before initial blood testing, 2 did not carry the pregnancy until initial blood testing, and 3 were lost to follow-up before initial blood testing. Thus, hemoglobin data were available for 173 patients (95%).

Thirty-eight women (22%) had anemia defined as a hemoglobin level lower than 11 mg per dL at any time during the pregnancy. One patient was excluded from further evaluation because she entered prenatal care late, had an initial hemoglobin level of 9.6 mg per dL at 39 weeks’ gestation, and was given iron without serum ferritin determination.

Of the 37 patients with anemia who had serum ferritin level measurements, the values ranged from 3 to 91 ng per dL. Twenty of these 37 patients (54%) had iron-deficiency anemia (ferritin levels ranging from 3-10 mg/dL), and 17 patients (46%) had anemia not related to iron deficiency (range=13-91 mg/dL).

Of the 38 patients with anemia by our definition, 20 had hemoglobin levels between 10.5 and 10.9 mg per dL, of whom 10 were between 14 and 28 weeks’ gestation (second trimester). Thus only 28 patients (16%) had anemia according to the CDC criteria. Of the 27 patients with anemia according to the CDC criteria (one patient was excluded), 17 (63%) had iron-deficiency anemia, and 10 (37%) had anemia not related to iron deficiency.

Discussion

Screening for anemia in pregnancy has been based on an association with an increased risk for preterm delivery.^1,2 This association is likely due to the comparison of the natural nadir in hemoglobin values in the second trimester, measured at the time of preterm delivery, with the higher hemoglobin values, measured at the time of term deliveries; this concept, however, has not been fully evaluated.¹ Despite the lack of evidence that screening for anemia improves clinical maternal, fetal, or neonatal outcomes¹⁰ standard obstetrical practice has been to screen all pregnant women for anemia and empirically treat anemic patients with iron therapy.

Although there is no convincing data demonstrating clear harm from iron therapy during pregnancy, limiting such therapy to patients with strictly defined anemia and demonstrated iron deficiency may be prudent to minimize potential harms from a practice not shown to clearly provide benefits.

In actual practice, with the common misdating of pregnancies and the potential inefficiency of complex management rules, it is possible that cutoff hemoglobin levels of 11 mg per dL for defining anemia in pregnancy without respect to gestational age are being used instead of the CDC criteria. By doing so, more patients without iron-deficiency anemia will be labeled as anemic.

The use of hematologic indices provided with complete blood count determination would be more efficient and less costly than serum ferritin determination, if they are shown to differentiate iron deficiency from other causes of anemia in pregnant women. Scatterplots have been used to estimate the discriminatory ability of continuous variables for discerning between patient populations.¹¹ We made scatterplots for RBC, MCV, MCH, MCHC, and RDW. None of these scatterplots suggested discriminatory ability for any of these variables. The results were unchanged when limited to patients with CDC-defined anemia.

Costs

Ferritin determination may be cost-effective depending on its cost, the cost of iron therapy, the prevalence of iron-deficiency anemia (which is dependent on the criteria for defining anemia), and the nonfinancial burden of unnecessary iron therapy. The costs of sparing women unnecessary iron therapy on the basis of these variables are detailed in the [Table]. Serum ferritin level determination cost $30 at the laboratory we used for our study. Other methods of determining iron deficiency (such as iron and total iron-binding capacity levels, which cost $35) were not evaluated, since the evidence did not suggest that these values were accurate or well known in pregnant women. Using numbers derived from our study, checking the ferritin levels of 100 pregnant women with anemia would cost $3000 and spare 37 to 46 women from taking iron, at a cost of $37.24 to $73.56 to prevent 1 woman from taking an unnecessary course of iron therapy.

Caretakers prescribing iron therapy are familiar with its adverse effects and relatively low tolerability. In a dose-finding study¹² of 110 pregnant women randomized to 1 of 3 doses of ferrous sulfate daily, 32.4% of those taking 60 mg of elemental iron (equivalent to 325 mg of ferrous sulfate) and 40.5% of those taking 120 mg of elemental iron (equivalent to common twice a day dosing) had side effects. Dropout rates matched the side effect rates (32.4% and 38.8%, respectively). Thus, for every 5 women treated with iron, 2 will develop side effects and stop taking it.

The cost-effectiveness of ferritin determination is highly dependent on and inversely related to the prevalence of iron-deficiency anemia in the patient population. As seen in the Table, if the prevalence of iron-deficiency anemia is sufficiently low, ferritin determination may be very cost-effective.

Clinicians should consider the local costs of ferrous sulfate, ferritin determination, and the prevalence of iron-deficiency anemia in their patient population in the evaluation of the use of ferritin determination instead of empiric iron therapy. Alternately, clinicians may present some of the issues and uncertainties to their patients for combined decision making.

Conclusions

In our population of prenatal patients with anemia, only 54% had an iron deficiency. Diagnostic and therapeutic approaches to screening for anemia in pregnancy should be reconsidered and further evaluated to avoid unnecessary iron therapy.

Acknowledgments

We would like to acknowledge Siobhan Ulrich, Jan Balmer, Sue Gibson, and Tracy Chappel for their efforts in implementing the practice protocol and keeping track of patient records.

Standard obstetrical practice has included screening for anemia and the provision of iron supplements to anemic patients. This approach has been based on assumptions about anemia and iron deficiency that are not supported by the literature.

Anemia in pregnancy has been reported to be associated with preterm delivery.^1,2 However, this may not take into consideration the lower hemoglobin values normally present during the second trimester. The potentially spurious association between anemia and preterm delivery could be explained by the lower hemoglobin values that are expected during the second trimester. When measured at the time of preterm delivery, these lower hemoglobin values, which are normal in the second trimester (but not in the third trimester or in a nonpregnant woman), are often believed to have led to the preterm delivery.

Normal hemogloblin values from nonpregnant women cannot be assumed to apply to those who are pregnant. Average hemoglobin levels decrease to 11.6 g per dL at 20 to 24 weeks’ gestation, with the fifth percentile at 10.5 g per dL (hematocrit=32%). Anemia in pregnancy has been defined by criteria from the Centers for Disease Control and Prevention (CDC) as a hemoglobin level of less than 11 g per dL during the first and third trimesters and less than 10.5 g per dL during the second trimester.³

When anemia is present in pregnancy it cannot be assumed to be the result of iron deficiency, even though this type of anemia has been previously reported as the most common cause.⁴ The Camden study⁵ of 826 pregnant women showed preterm delivery and low birth weight associated with iron-deficiency anemia. Only 27.9% of the pregnant women had anemia, however, and only 12.5% of the patients with anemia had an iron deficiency. Thus, only 3.5% of the entire cohort had iron-deficiency anemia.⁶

Iron deficiency in pregnancy has been defined by the National Academy of Sciences panel on nutrition and pregnancy⁷ as ferritin levels lower than 12 ng per mL. A systematic overview⁸ of 55 studies relevant to laboratory tests for diagnosis of iron-deficiency anemia in variable patient populations found serum ferritin radioimmunoassay to be the most powerful test. Ferritin levels are considered the gold standard for the diagnosis of iron-deficiency anemia in pregnancy.⁹

We report a descriptive study of the use of ferritin levels to determine the need for iron supplementation among pregnant women with anemia.

Methods

Lebanon Family Health Services is a nonprofit federally subsidized community agency providing prenatal care and women’s health services to a diverse population without restriction on the basis of financial status. The prenatal patients cared for in this practice included women aged 15 to 40 years (23.6% were younger than 19 years; 70.8%, 19 to 30 years; and 5.6%, older than 30 years) of whom 65.3% were white, 29.2% Hispanic, 4.2% African American, and 1.4% Asian.

We evaluated all patients entering into prenatal care at Lebanon Family Health Services from April 1, 1997, through December 31, 1998, using prospective data collection and retrospective record review. Prenatal vitamins (including elemental iron 30 to 60 mg/day) were prescribed to all patients. Complete blood count was tested at the initial evaluation as part of a comprehensive screening. For patients who entered prenatal care at earlier than 28 weeks’ gestation complete blood count was checked again when they had reached that point.

The CDC criteria for defining anemia are hemoglobin levels less than 11 g per dL during the first and third trimesters, and less than 10.5 g per dL during the second trimester. A hemoglobin level of less than 11 g per dL at any time during the pregnancy was used as the cutoff point for anemia in this clinical practice, in consideration of patients with uncertain or inaccurate pregnancy dating.

Patients with anemia underwent blood testing for serum ferritin level, generally 1 week after a complete blood cell count. At the time, ferritin levels were determined using different test tubes than those used for other prenatal testing, thus resulting in a delay in obtaining blood for ferritin testing. If serum ferritin was 12 ng per dL or lower, iron-deficiency anemia was diagnosed and ferrous sulfate was prescribed for the remainder of the pregnancy and the postpartum period. If serum ferritin was greater than 12 ng per dL iron deficiency was excluded, anemia was generally considered to be pregnancy related, and further evaluation and treatment was at the discretion of the treating clinician.

Variables recorded for all patients with anemia were: estimated date of delivery, last menstrual period, date of any testing for complete blood count or ferritin levels, hemoglobin value, hematocrit value, red blood cell count (RBC), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), red cell distribution width (RDW), any notation regarding microcytosis, and serum ferritin levels.

Results

A total of 182 patients were consecutively entered into prenatal care during the study period. Hemoglobin data were not obtained from 9 patients—4 transferred to another practice for care before initial blood testing, 2 did not carry the pregnancy until initial blood testing, and 3 were lost to follow-up before initial blood testing. Thus, hemoglobin data were available for 173 patients (95%).

Thirty-eight women (22%) had anemia defined as a hemoglobin level lower than 11 mg per dL at any time during the pregnancy. One patient was excluded from further evaluation because she entered prenatal care late, had an initial hemoglobin level of 9.6 mg per dL at 39 weeks’ gestation, and was given iron without serum ferritin determination.

Of the 37 patients with anemia who had serum ferritin level measurements, the values ranged from 3 to 91 ng per dL. Twenty of these 37 patients (54%) had iron-deficiency anemia (ferritin levels ranging from 3-10 mg/dL), and 17 patients (46%) had anemia not related to iron deficiency (range=13-91 mg/dL).

Of the 38 patients with anemia by our definition, 20 had hemoglobin levels between 10.5 and 10.9 mg per dL, of whom 10 were between 14 and 28 weeks’ gestation (second trimester). Thus only 28 patients (16%) had anemia according to the CDC criteria. Of the 27 patients with anemia according to the CDC criteria (one patient was excluded), 17 (63%) had iron-deficiency anemia, and 10 (37%) had anemia not related to iron deficiency.

Discussion

Screening for anemia in pregnancy has been based on an association with an increased risk for preterm delivery.^1,2 This association is likely due to the comparison of the natural nadir in hemoglobin values in the second trimester, measured at the time of preterm delivery, with the higher hemoglobin values, measured at the time of term deliveries; this concept, however, has not been fully evaluated.¹ Despite the lack of evidence that screening for anemia improves clinical maternal, fetal, or neonatal outcomes¹⁰ standard obstetrical practice has been to screen all pregnant women for anemia and empirically treat anemic patients with iron therapy.

Although there is no convincing data demonstrating clear harm from iron therapy during pregnancy, limiting such therapy to patients with strictly defined anemia and demonstrated iron deficiency may be prudent to minimize potential harms from a practice not shown to clearly provide benefits.

In actual practice, with the common misdating of pregnancies and the potential inefficiency of complex management rules, it is possible that cutoff hemoglobin levels of 11 mg per dL for defining anemia in pregnancy without respect to gestational age are being used instead of the CDC criteria. By doing so, more patients without iron-deficiency anemia will be labeled as anemic.

The use of hematologic indices provided with complete blood count determination would be more efficient and less costly than serum ferritin determination, if they are shown to differentiate iron deficiency from other causes of anemia in pregnant women. Scatterplots have been used to estimate the discriminatory ability of continuous variables for discerning between patient populations.¹¹ We made scatterplots for RBC, MCV, MCH, MCHC, and RDW. None of these scatterplots suggested discriminatory ability for any of these variables. The results were unchanged when limited to patients with CDC-defined anemia.

Costs

Ferritin determination may be cost-effective depending on its cost, the cost of iron therapy, the prevalence of iron-deficiency anemia (which is dependent on the criteria for defining anemia), and the nonfinancial burden of unnecessary iron therapy. The costs of sparing women unnecessary iron therapy on the basis of these variables are detailed in the [Table]. Serum ferritin level determination cost $30 at the laboratory we used for our study. Other methods of determining iron deficiency (such as iron and total iron-binding capacity levels, which cost $35) were not evaluated, since the evidence did not suggest that these values were accurate or well known in pregnant women. Using numbers derived from our study, checking the ferritin levels of 100 pregnant women with anemia would cost $3000 and spare 37 to 46 women from taking iron, at a cost of $37.24 to $73.56 to prevent 1 woman from taking an unnecessary course of iron therapy.

Caretakers prescribing iron therapy are familiar with its adverse effects and relatively low tolerability. In a dose-finding study¹² of 110 pregnant women randomized to 1 of 3 doses of ferrous sulfate daily, 32.4% of those taking 60 mg of elemental iron (equivalent to 325 mg of ferrous sulfate) and 40.5% of those taking 120 mg of elemental iron (equivalent to common twice a day dosing) had side effects. Dropout rates matched the side effect rates (32.4% and 38.8%, respectively). Thus, for every 5 women treated with iron, 2 will develop side effects and stop taking it.

The cost-effectiveness of ferritin determination is highly dependent on and inversely related to the prevalence of iron-deficiency anemia in the patient population. As seen in the Table, if the prevalence of iron-deficiency anemia is sufficiently low, ferritin determination may be very cost-effective.

Clinicians should consider the local costs of ferrous sulfate, ferritin determination, and the prevalence of iron-deficiency anemia in their patient population in the evaluation of the use of ferritin determination instead of empiric iron therapy. Alternately, clinicians may present some of the issues and uncertainties to their patients for combined decision making.

Conclusions

In our population of prenatal patients with anemia, only 54% had an iron deficiency. Diagnostic and therapeutic approaches to screening for anemia in pregnancy should be reconsidered and further evaluated to avoid unnecessary iron therapy.

Acknowledgments

We would like to acknowledge Siobhan Ulrich, Jan Balmer, Sue Gibson, and Tracy Chappel for their efforts in implementing the practice protocol and keeping track of patient records.

Standard obstetrical practice has included screening for anemia and the provision of iron supplements to anemic patients. This approach has been based on assumptions about anemia and iron deficiency that are not supported by the literature.

Anemia in pregnancy has been reported to be associated with preterm delivery.^1,2 However, this may not take into consideration the lower hemoglobin values normally present during the second trimester. The potentially spurious association between anemia and preterm delivery could be explained by the lower hemoglobin values that are expected during the second trimester. When measured at the time of preterm delivery, these lower hemoglobin values, which are normal in the second trimester (but not in the third trimester or in a nonpregnant woman), are often believed to have led to the preterm delivery.

Normal hemogloblin values from nonpregnant women cannot be assumed to apply to those who are pregnant. Average hemoglobin levels decrease to 11.6 g per dL at 20 to 24 weeks’ gestation, with the fifth percentile at 10.5 g per dL (hematocrit=32%). Anemia in pregnancy has been defined by criteria from the Centers for Disease Control and Prevention (CDC) as a hemoglobin level of less than 11 g per dL during the first and third trimesters and less than 10.5 g per dL during the second trimester.³

When anemia is present in pregnancy it cannot be assumed to be the result of iron deficiency, even though this type of anemia has been previously reported as the most common cause.⁴ The Camden study⁵ of 826 pregnant women showed preterm delivery and low birth weight associated with iron-deficiency anemia. Only 27.9% of the pregnant women had anemia, however, and only 12.5% of the patients with anemia had an iron deficiency. Thus, only 3.5% of the entire cohort had iron-deficiency anemia.⁶

Iron deficiency in pregnancy has been defined by the National Academy of Sciences panel on nutrition and pregnancy⁷ as ferritin levels lower than 12 ng per mL. A systematic overview⁸ of 55 studies relevant to laboratory tests for diagnosis of iron-deficiency anemia in variable patient populations found serum ferritin radioimmunoassay to be the most powerful test. Ferritin levels are considered the gold standard for the diagnosis of iron-deficiency anemia in pregnancy.⁹

We report a descriptive study of the use of ferritin levels to determine the need for iron supplementation among pregnant women with anemia.

Methods

Lebanon Family Health Services is a nonprofit federally subsidized community agency providing prenatal care and women’s health services to a diverse population without restriction on the basis of financial status. The prenatal patients cared for in this practice included women aged 15 to 40 years (23.6% were younger than 19 years; 70.8%, 19 to 30 years; and 5.6%, older than 30 years) of whom 65.3% were white, 29.2% Hispanic, 4.2% African American, and 1.4% Asian.

We evaluated all patients entering into prenatal care at Lebanon Family Health Services from April 1, 1997, through December 31, 1998, using prospective data collection and retrospective record review. Prenatal vitamins (including elemental iron 30 to 60 mg/day) were prescribed to all patients. Complete blood count was tested at the initial evaluation as part of a comprehensive screening. For patients who entered prenatal care at earlier than 28 weeks’ gestation complete blood count was checked again when they had reached that point.

The CDC criteria for defining anemia are hemoglobin levels less than 11 g per dL during the first and third trimesters, and less than 10.5 g per dL during the second trimester. A hemoglobin level of less than 11 g per dL at any time during the pregnancy was used as the cutoff point for anemia in this clinical practice, in consideration of patients with uncertain or inaccurate pregnancy dating.

Patients with anemia underwent blood testing for serum ferritin level, generally 1 week after a complete blood cell count. At the time, ferritin levels were determined using different test tubes than those used for other prenatal testing, thus resulting in a delay in obtaining blood for ferritin testing. If serum ferritin was 12 ng per dL or lower, iron-deficiency anemia was diagnosed and ferrous sulfate was prescribed for the remainder of the pregnancy and the postpartum period. If serum ferritin was greater than 12 ng per dL iron deficiency was excluded, anemia was generally considered to be pregnancy related, and further evaluation and treatment was at the discretion of the treating clinician.

Variables recorded for all patients with anemia were: estimated date of delivery, last menstrual period, date of any testing for complete blood count or ferritin levels, hemoglobin value, hematocrit value, red blood cell count (RBC), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), red cell distribution width (RDW), any notation regarding microcytosis, and serum ferritin levels.

Results

A total of 182 patients were consecutively entered into prenatal care during the study period. Hemoglobin data were not obtained from 9 patients—4 transferred to another practice for care before initial blood testing, 2 did not carry the pregnancy until initial blood testing, and 3 were lost to follow-up before initial blood testing. Thus, hemoglobin data were available for 173 patients (95%).

Thirty-eight women (22%) had anemia defined as a hemoglobin level lower than 11 mg per dL at any time during the pregnancy. One patient was excluded from further evaluation because she entered prenatal care late, had an initial hemoglobin level of 9.6 mg per dL at 39 weeks’ gestation, and was given iron without serum ferritin determination.

Of the 37 patients with anemia who had serum ferritin level measurements, the values ranged from 3 to 91 ng per dL. Twenty of these 37 patients (54%) had iron-deficiency anemia (ferritin levels ranging from 3-10 mg/dL), and 17 patients (46%) had anemia not related to iron deficiency (range=13-91 mg/dL).

Of the 38 patients with anemia by our definition, 20 had hemoglobin levels between 10.5 and 10.9 mg per dL, of whom 10 were between 14 and 28 weeks’ gestation (second trimester). Thus only 28 patients (16%) had anemia according to the CDC criteria. Of the 27 patients with anemia according to the CDC criteria (one patient was excluded), 17 (63%) had iron-deficiency anemia, and 10 (37%) had anemia not related to iron deficiency.

Discussion

Screening for anemia in pregnancy has been based on an association with an increased risk for preterm delivery.^1,2 This association is likely due to the comparison of the natural nadir in hemoglobin values in the second trimester, measured at the time of preterm delivery, with the higher hemoglobin values, measured at the time of term deliveries; this concept, however, has not been fully evaluated.¹ Despite the lack of evidence that screening for anemia improves clinical maternal, fetal, or neonatal outcomes¹⁰ standard obstetrical practice has been to screen all pregnant women for anemia and empirically treat anemic patients with iron therapy.

Although there is no convincing data demonstrating clear harm from iron therapy during pregnancy, limiting such therapy to patients with strictly defined anemia and demonstrated iron deficiency may be prudent to minimize potential harms from a practice not shown to clearly provide benefits.

In actual practice, with the common misdating of pregnancies and the potential inefficiency of complex management rules, it is possible that cutoff hemoglobin levels of 11 mg per dL for defining anemia in pregnancy without respect to gestational age are being used instead of the CDC criteria. By doing so, more patients without iron-deficiency anemia will be labeled as anemic.

The use of hematologic indices provided with complete blood count determination would be more efficient and less costly than serum ferritin determination, if they are shown to differentiate iron deficiency from other causes of anemia in pregnant women. Scatterplots have been used to estimate the discriminatory ability of continuous variables for discerning between patient populations.¹¹ We made scatterplots for RBC, MCV, MCH, MCHC, and RDW. None of these scatterplots suggested discriminatory ability for any of these variables. The results were unchanged when limited to patients with CDC-defined anemia.

Costs

Ferritin determination may be cost-effective depending on its cost, the cost of iron therapy, the prevalence of iron-deficiency anemia (which is dependent on the criteria for defining anemia), and the nonfinancial burden of unnecessary iron therapy. The costs of sparing women unnecessary iron therapy on the basis of these variables are detailed in the [Table]. Serum ferritin level determination cost $30 at the laboratory we used for our study. Other methods of determining iron deficiency (such as iron and total iron-binding capacity levels, which cost $35) were not evaluated, since the evidence did not suggest that these values were accurate or well known in pregnant women. Using numbers derived from our study, checking the ferritin levels of 100 pregnant women with anemia would cost $3000 and spare 37 to 46 women from taking iron, at a cost of $37.24 to $73.56 to prevent 1 woman from taking an unnecessary course of iron therapy.

Caretakers prescribing iron therapy are familiar with its adverse effects and relatively low tolerability. In a dose-finding study¹² of 110 pregnant women randomized to 1 of 3 doses of ferrous sulfate daily, 32.4% of those taking 60 mg of elemental iron (equivalent to 325 mg of ferrous sulfate) and 40.5% of those taking 120 mg of elemental iron (equivalent to common twice a day dosing) had side effects. Dropout rates matched the side effect rates (32.4% and 38.8%, respectively). Thus, for every 5 women treated with iron, 2 will develop side effects and stop taking it.

The cost-effectiveness of ferritin determination is highly dependent on and inversely related to the prevalence of iron-deficiency anemia in the patient population. As seen in the Table, if the prevalence of iron-deficiency anemia is sufficiently low, ferritin determination may be very cost-effective.

Clinicians should consider the local costs of ferrous sulfate, ferritin determination, and the prevalence of iron-deficiency anemia in their patient population in the evaluation of the use of ferritin determination instead of empiric iron therapy. Alternately, clinicians may present some of the issues and uncertainties to their patients for combined decision making.

Conclusions

In our population of prenatal patients with anemia, only 54% had an iron deficiency. Diagnostic and therapeutic approaches to screening for anemia in pregnancy should be reconsidered and further evaluated to avoid unnecessary iron therapy.

Acknowledgments

We would like to acknowledge Siobhan Ulrich, Jan Balmer, Sue Gibson, and Tracy Chappel for their efforts in implementing the practice protocol and keeping track of patient records.

Postherpetic neuralgia (PHN) is the most common complication of zoster and is much more prevalent among older patients.^1,2 The results of the largest English-language prospective study of patients presenting with zoster,¹ involving 457 patients from 62 general practitioners in Iceland, suggest that the average family physician caring for 2000 patients would expect to see 4 cases of zoster per year and one case of PHN lasting more than 3 months once in 3 years. Among patients older than 60 years in this study, 19% had pain at 3 months, and 8% had pain at 12 months. Among the 183 patients younger than 40 years, only one had pain at 3 months. The authors of the second largest prospective study of zoster,² involving 206 patients from multiple specialty services in Philadelphia, Pennsylvania, found that 14% of patients older than 50 years had pain at 3 months, and 7% had pain at 6 months. In this study, no patients younger than 50 years had pain at 3 months. In comparison, randomized controlled trials of zoster treatment have reported ranges of 16.7% to 60% of patients receiving placebo having pain at 3 months and 5% to 39.1% having pain at 6 months.

Several systematic reviews have addressed treatment or prevention of PHN.^3-5 New trial data^6-9 and discrepancies in the literature^4,10,11 prompted our paper. Differences in the definition of PHN complicated our review of the medical literature.^12,13 For this review, PHN is defined as any pain after cutaneous healing of zoster, and zoster-associated pain (ZAP) is defined as any pain associated with zoster (acute zoster pain and PHN). This review is limited to randomized controlled trials performed at the time of zoster with follow-up addressing the incidence or duration of PHN or ZAP.

Methods

MEDLINE (1966 to present) was searched on December 29, 1998, using PubMed and combining the terms “zoster” or “post-herpetic neuralgia” or “postherpetic neuralgia” or “post-herpetic pain” or “postherpetic pain,” and publication type “clinical trial” (including phases I-IV) or “controlled clinical trial” or “randomized controlled trial.” We searched The Cochrane Controlled Trials Register 1998, issue 4, using the same terms. We also identified trials through article reference lists and Web-based searches. One author conducted independent searches and was unable to identify any additional randomized controlled trials.

For our review we included randomized controlled trials published in English that enrolled primarily immunocompetent subjects with acute herpes zoster and addressed relevant end points (incidence of pain at any time after rash healing in zoster patients or duration of ZAP or PHN). The two authors independently evaluated the trials meeting these inclusion criteria for quality of randomization, allocation concealment, blinding, baseline difference assessment, methods of data collection, adequacy of follow-up (duration and methods), accounting for dropouts, and intention-to-treat analysis. We rated methodologic quality as good, fair, or poor on the basis of an overall assessment of these features. We did not use explicit validity checklists with summary scores, because they have not been shown to predict the effect of bias on treatment differences (ie, they have not been shown to provide more reliable assessments of validity).^14,15 Discordant ratings of trial quality were resolved through the consensus of both investigators.

Results

Our literature search identified 74 appropriate trials. Of the 74 trials reviewed, 42 pertained to treatment at the time of acute zoster to prevent PHN, and 32 pertained to treatment of established PHN.* Of the 42 prevention trials reviewed, 6 were rated as good quality, 22 as fair, and 14 as poor.

Acyclovir

Nucleoside analogues represent the mainstay of acute herpes zoster treatment and produce a faster rate of cutaneous healing and decreased risk of ophthalmic complications.^16,17 A limited subset of acute zoster treatment trials have reported data on PHN incidence or duration.

Four placebo-controlled trials have been conducted using oral acyclovir 800 mg 5 times daily for 7 to 10 days within 72 hours of zoster rash (Table 1).

The largest placebo-controlled trial of acyclovir involved 376 patients from 3 United Kingdom centers and has been reported in multiple publications.^16,18-22 Acyclovir had no effect on the incidence or severity of PHN. Follow-up was monthly to 6 months or until there was no pain during the previous month; this loss to follow-up after the first cessation of pain is the major criticism of this trial, because PHN is known to relapse.^22,23 With this loss to follow-up, the amount of time to complete cessation of pain could not be determined.^23,24 Ten-year follow-up of 132 patients from this trial who reported resolution of pain found that pain returned in 16 cases (12%).²² In subsequent publications, investigators have reported conflicting results in terms of the long-term incidence of PHN. A statistically significant difference in the incidence of PHN was reported in 57 patients from Southampton contacted at 5 years (7% acyclovir patients vs 37% placebo patients, P = .01),²¹ but not in 160 patients from the Sheffield and Birmingham centers contacted at 9 to 10 years.²²

A smaller US multicenter placebo-controlled trial of oral acyclovir reported decreased pain at 1 to 3 months posttreatment, but no differences in analgesic use were reported.^25,26 A reanalysis of 166 patients with pain at enrollment found a reduction in the median duration of ZAP from 62 days to 20 days with acyclovir (P = .02).²⁷ However, of the 21 patients with no pain on enrollment, 15 later developed pain and 4 of those patients continued to report pain through the seventh month. It is possible that inclusion of these patients in the analysis would materially change the results.

In a third trial²⁸ consisting of 83 patients presenting to general practitioners, acyclovir was associated with statistically significant less pain to 3 months. Total analgesic use, however, was reduced in the acyclovir group for only the first 4 weeks.

The fourth trial,²⁹ consisting of 46 patients with acute herpes zoster ophthalmicus, acyclovir was associated with a reduction in pain incidence that was statistically significant at 2 months (P = .04), but not at 6 months (P = .07). Pain severity was significantly reduced from 2 through 6 months, although this analysis is based on only 6 patients having pain at 6 months.

These 4 trials provided all or most of the data for several reviews and meta-analyses.^4,23,24,30 Summary results among these reviews varied because of different methods used, different subgroups assessed, and different efficacy end points reported but generally supported short-term benefit for oral acyclovir. A key issue affecting the overall results of these reviews was consideration of time to complete cessation of pain. This end point was clearly recorded in the 3 smaller trials with positive results but not reported in the large trial with negative results as discussed above.

Crooks and colleagues²³ performed a pooled analysis and reported that oral acyclovir reduced the incidence of PHN from 19% to 11% at 3 to 6 months. We (as well as Lancaster and colleagues⁴) were unable to verify the placebo group incidence rates reported by Crooks and coworkers which appear to have overestimated any acyclovir benefit, and therefore cannot substantiate these findings. Analysis for time to first cessation of pain was reported as a nonsignificant trend toward benefit with acyclovir, while analysis for time to complete cessation of pain (which excluded the largest trial with the negative results) was reported as a statistically significant reduction from an average 86 to 49 days.

Lancaster and colleagues⁴ found a statistically significant reduction in pain at 3 months but no statistically significant benefit at 1 month or 6 months. They commented that analgesic use was not significantly different in 2 of the positive studies.

Wood and coworkers²⁴ derived results from previously unpublished data from these 4 trials and reported that acyclovir significantly accelerated pain resolution. Significant reductions in incidence of pain with acyclovir were reported at 3 and 6 months, but it is unclear if the largest of the 4 trials was included in this analysis.

In the most recent meta-analysis,³⁰ Jackson and colleagues included a fifth trial³¹ and reported an absolute risk reduction of 16% for the incidence of any pain at 6 months (P <.05; number needed to treat [NNT] = 6.3). Multiple methodologic flaws in that meta-analysis, detailed in the Cochrane Review Database,³² make application of the results unclear.

Additional oral acyclovir trials are summarized in Table 1.* Extending the duration of treatment to 14 or 21 days did not provide benefit over 7 days.^33,34 Lower doses of acyclovir had no effect on incidence, severity, or duration of PHN.^17,35,36 There is no evidence to support use of intravenous or topical acyclovir for prevention of PHN.*

Marginal evidence exists to suggest that oral acyclovir 800 mg 5 times daily may reduce the incidence of pain at 1 to 3 months. Information regarding effects on quality of life were generally lacking.

Newer Antivirals

Famciclovir, a prodrug of penciclovir, did not affect the incidence of PHN measured at the time of cutaneous healing in a placebo-controlled trial of 419 patients6 (Table 2). Among the 186 patients (44%) who developed PHN, famciclovir significantly reduced its duration by a median of 2 months, 3.5 months in patients older than 50 years. No dose-response difference was noted. In a subsequent publication presenting monthly prevalence data,³⁷ pain 6 months after enrollment was reported in 15% of patients assigned famciclovir 500 mg and 23.8% of the patients assigned placebo (NNT = 11.4). Similar results were stated for the 750-mg group. Results were reported as statistically significant but P values were not given.

Valacyclovir, a prodrug of acyclovir, has not been studied in a placebo-controlled trial in patients older than 50 years, based on published reports. In a comparison trial with acyclovir,⁷ valacyclovir for 7 or 14 days did not reduce the incidence of PHN (pain after rash healing) but did accelerate its resolution by 1 to 2 weeks. Pain persisting for 6 months was found in 25.7% of the acyclovir group and 18.6% of the combined valacyclovir group (P = .02; NNT = 15.6). The actual benefit of valacyclovir cannot be determined, since the evidence for acyclovir is inconclusive.

Steroids

Inflammation of peripheral nervous system structures has been identified in specimens from patients with PHN.³⁸ Because of this, corticosteroids have been used in the treatment of herpes zoster in hopes of preventing PHN. Trials evaluating steroids are heterogeneous, involving different drugs, doses, routes, durations, and follow-up methods, thus hampering pooling of trial data (Table 3).

Esmann and coworkers³⁹ compared a 21-day oral prednisolone treatment with placebo in 84 patients, all of whom received acyclovir for the first 7 days. Pretrial calculation demonstrated that 324 patients would be required to detect an 80% change in incidence of PHN at 6 months with statistical significance. Enrollment was stopped early when an interim analysis did not show any treatment effect at 3 months at the P <.10 level. No benefit was seen at 6 months. Data points before 6 months were not reported. This is the methodologically strongest of the 4 early steroid trials.

Clemmensen and Andersen⁴⁰ randomized 60 patients to adrenocorticotropic hormone, prednisone, or placebo. No benefit was found for prednisone, and both active treatments may have increased pain at 1 month. No data were reported beyond 6 weeks.

Keczkes and Basheer⁴¹ randomized 40 patients with “severe, painful” zoster to prednisolone or carbamazepine for 4 weeks. A decreased incidence in pain was reported with prednisolone at 8 weeks, but statistics were not reported. Of the 4 early steroid trials, this shows the greatest benefit but is methodologically the weakest study. Without a placebo group, the purported benefit of prednisolone due to a detrimental effect of carbamazepine cannot be excluded.

Eaglstein and colleagues⁴² randomized 35 patients with “severely painful” zoster to triamcinolone versus placebo. No patients younger than 60 years developed neuralgia. Pain reduction with triamcinolone was reported at 8 weeks but not at 6 months for the 24 patients who developed neuralgia. Statistics were not reported.

These 4 trials have undergone several reviews. In a meta-analysis, Lycka10 reported statistically significant reductions in pain at 6 and 12 weeks after zoster onset but not at 6 months. That meta-analysis incorporated unpublished data obtained from the original investigators. NNTs of 4.5 to prevent pain at 6 weeks and 3.5 to prevent pain at 12 weeks were reported, although an intention-to-treat analysis was not performed. The authors of 2 other reviews^4,11 concluded that there was insufficient evidence to support a benefit from steroids, finding the trials too heterogeneous to appropriately combine in meta-analysis.⁴

A later placebo-controlled trial³¹ randomized 201 patients to acyclovir, prednisone, both, or neither for 21 days. That trial had good methodology and may have been large enough to provide the evidence to support or refute the benefit of steroids. No differences in pain incidence were found at 3 or 6 months. Benefit during the first month was found, but results were reported only as risk reductions.With no incidence or duration data reported, these results are not clinically applicable. Follow-up commentary⁴³ pointed out the need for actual data instead of risk ratios. The authors’ response provided limited data showing reductions in median time to cessation of analgesic use (from 28 to 14 days), return to normal activity (from 21 to 3 days), and uninterrupted sleep (from 26 to 5 days) for the combined acyclovir and prednisone treatment group compared with the placebo group. No data were provided for the acyclovir-only or prednisone-only groups.

Wood and colleagues³⁴ randomized 400 patients to prednisolone or placebo for 21 days. Patients were also randomized to acyclovir for 7 days or 21 days, so there was no placebo-only group. There was no benefit to extending acyclovir to 21 days or to adding prednisolone with respect to incidence or duration of PHN.

A recent review⁴⁴ incorporating these newer trials concluded that steroids can reduce acute herpes zoster pain and improve short-term (1-month) quality of life and suggested that steroids are reasonable to use in patients older than 50 years. No effect, however, was demonstrated on the incidence, severity, or duration of PHN.

Tricyclic Antidepressants

Amitriptyline exerts a pain-modulating effect separate from its antidepressant properties⁴⁵ and has been widely used in neuropathic and other chronic pain states. Amitriptyline (25 mg nightly) used preemptively for 90 days starting within 48 hours of rash onset showed a statistically significant reduction in pain incidence at 6 months in a single placebo-controlled trial of fair quality.⁸ Blinding may have been inadequate, because patients were warned of potential dry mouth. Follow-up was a single contact by telephone or mail at 6 to 8 months. Acyclovir given by general practitioners was not controlled. Twenty-four percent of the amitriptyline group and 50% of the placebo group received acyclovir. Among amitriptyline-treated patients, there was a nonsignificant trend toward reduced pain with acyclovir use, while placebo patients who received acyclovir experienced a nonsignificant trend toward increased pain.

Percutaneous Electrical Nerve Stimulation

Percutaneous electrical nerve stimulation (PENS) compared favorably with famciclovir in a single blind randomized trial of 50 adults with zoster.⁴⁶ A 12% absolute risk reduction in pain was reported at 3 and 6 months (but not at 9 months), but no statistical analysis was reported for this measure. Statistically significant reductions in severity of pain were found at 3 and 6 months. Costs and availability of PENS were not reported.

Discussion

There are limited data from randomized controlled trials that indicate that early treatment of acute herpes zoster decreases the incidence or duration of PHN. Acyclovir was the most-studied agent, but there is no convincing evidence that acyclovir alters the course of PHN. There is some evidence that oral acyclovir 800 mg 5 times daily for 7 to 10 days reduces the incidence of pain in the short term (1 to 3 months). Valacyclovir was somewhat more effective than acyclovir in the single largest antiviral trial, but without a placebo control, the actual efficacy of either drug is indeterminate. Famciclovir did not alter the incidence of PHN (at the time of rash healing) but did significantly reduce the duration of PHN in a single placebo-controlled trial of good methodology.

Randomized antiviral trials have been limited totrials including patients presenting within 72 hours of rash onset. No data are available to address the use of antivirals initiated more than 72 hours after rash onset. In the largest prospective study of patients presenting to general practitioners with zoster,¹ only 44% presented within 72 hours of rash onset.

For immunocompetent subjects, oral acyclovir, famciclovir, and valacyclovir were free of major toxicities and demonstrated side effects comparable with placebo in the clinical trials presented. According to the package inserts, posttrial case reports of more serious reactions, including anaphylaxis and renal failure, have been cited. These medications require dose adjustment in patients with compromised renal function.

Steroids, like antivirals, are widely prescribed for the treatment of acute herpes zoster. Although useful for reduction of early pain, there is no evidence that systemic steroids prevent or shorten the course of PHN. No further trials are needed in this regard. The largest and best designed trial involving corticosteroids for the treatment of zoster is rendered clinically inapplicable by the presentation of relative risk reduction rather than incidence and duration data for placebo and treatment groups.

Preemptive amitriptyline and PENS appear promising on the basis of single trials of fair quality. Potentially promising agents based on poor quality trials* include amantadine and Clinacanthus nutans cream. On the basis of the quality of available evidence, availability of therapy, and costs (Table 5), amitriptyline is the most promising of these agents. Anticholinergic side effects are likely to be of greatest significance in the population at risk for PHN (the elderly), although only 3 of 41 patients (7%) withdrew from therapy in the amitriptyline trial.⁸

Limitations

A systematic review of PHN is hampered by different definitions of PHN (ranging from pain immediately following rash healing to 6 months after rash onset), differences in primary end points measured, and differences in study follow-up methods and duration. Although we stated our definition of PHN (pain following rash healing), we have not restricted our analysis to studies employing the same definition or attempted to analyze only study data using that definition. There is debate about whether a pain continuum (ZAP) or subdivision (acute, subacute, and chronic) of zoster-related pain is most suitable for randomized controlled trials examining the impact of treatments on PHN.^47,48 Although the pain that often heralds and more frequently accompanies zoster may merit treatment, we are particularly interested in whether such treatment will have an impact on the pain incidence, duration, and severity that follow rash healing.

We used a comprehensive literature-based search. Searches for unpublished literature and contact with investigators were generally not undertaken. There is a strong chance of publication bias because of our methods of searching. For example, the product information for valacyclovir notes a placebo-controlled trial in patients younger than 50 years presenting with 72 hours of zoster that found no difference with respect to the duration of pain after rash healing. We were unable to locate this trial in a published format, so it did not meet our inclusion criteria.

Language bias is another source of publication basis. We only included studies published in English. Sixteen non–English-language studies identified were potentially pertinent to the prevention of PHN.*

Quality-of-life measures may be ultimately more important than measures of incidence or duration for assessing the impact of treatment on patients. Few trials^7,31,34,46 addressed quality-of-life measures, and these were generally reported only for the short term. Even in the newer antirviral trials demonstrating reduction in the duration of PHN,^6,7 a significant impact on quality of life was not well documented.

Analysis of the power of trials with negative results was not performed, so potential benefits of treatments not studied adequately cannot be excluded.

The expected outcome without treatment is an important consideration in decisions regarding prevention of PHN. There is an inherent selection bias in randomized controlled trials because an unknown number of patients with zoster will not present for medical attention. There is also a reporting bias with identification of PHN in subjects who would otherwise not be troubled enough by symptoms to present for medical care. Therefore, natural history data derived from placebo cohorts of randomized controlled trials is likely to overestimate the true incidence of clinically significant PHN.

Recommendations for future research

Trials of patients with acute herpes zoster should include all the standard criteria for good methodology (adequate randomization, blinding, and so forth), adequate numbers to detect significant differences determined a priori, continued follow-up of all randomized patients for at least 6 months, and detailed descriptions of the studied population. Trials should be limited to subjects older than 50 years, since this is the age group most likely to be afflicted with PHN. Trials should also evaluate patient-oriented end points other than pain, such as quality of life and time to return to usual activity, and should consider analgesic use as a surrogate pain measure. Cost, compliance, and tolerability should be assessed. For clinical applications, P values and NNT should be reported.

A comparison trial of famciclovir and valacyclovir is recommended. Future use of placebo arms in zoster trials is a matter of continued debate,^34,49 especially for trials enrolling patients with ophthalmic zoster.

Future trials of amitriptyline should extend enrollment to 96 hours or more. Inclusion in the amitriptyline trial was limited to subjects presenting within 48 hours of rash onset. This limitation would be very restrictive in clinical practice.

Further investigation of PENS and possibly Clinacanthus nutans cream and amantadine should be considered.

Primary prevention of PHN may best be achieved through prevention of varicella and subsequent zoster through vaccinations. It has been demonstrated in 2 immunized populations (children with leukemia and renal transplant patients) that zoster occurs 5 to 7 times less frequently after vaccination than after natural varicella.⁵⁰ Follow-up of immunocompromised and immunocompetent vaccine recipients is warranted to evaluate the vaccine’s impact on the natural history of zoster and PHN. Waning cell-mediated immunity to varicella-zoster virus with age has been associated with zoster, and booster vaccinations in the elderly have been shown to improve laboratory markers of cell-mediated immunity. Thus, there is a possibility that vaccinations of elderly patients may reduce the impact of zoster and subsequent PHN.⁵¹ Such a trial is currently recruiting 37,000 volunteers older than 60 years who have had chickenpox but have never had shingles.⁵² It should be noted that theoretical arguments have been made that primary vaccination of varicella could lead to subsequent increases in zoster incidence.⁵³

Recommendations for clinical practice

When acute herpes zoster occurs, providers should restrict their attention regarding PHN prevention to patients older than 50 years. According to the available evidence, antiviral therapy (famciclovir or valacyclovir for 7 days) started within 72 hours of rash onset and/or low-dose amitriptyline (for 90 days) may be offered in an effort to reduce PHN incidence or duration. Patients should be informed of the natural history of zoster and PHN, expected benefits for medications, potential for adverse effects, and expected costs. Other treatments, such as steroids or analgesics, may be offered to alter the acute course of zoster but would have no effect on preventing PHN.

Acknowledgments

The authors wish to thank Dehorah Lovett, Karen J. Alper, and Stacey Raulzhan for their assistance.

Postherpetic neuralgia (PHN) is the most common complication of zoster and is much more prevalent among older patients.^1,2 The results of the largest English-language prospective study of patients presenting with zoster,¹ involving 457 patients from 62 general practitioners in Iceland, suggest that the average family physician caring for 2000 patients would expect to see 4 cases of zoster per year and one case of PHN lasting more than 3 months once in 3 years. Among patients older than 60 years in this study, 19% had pain at 3 months, and 8% had pain at 12 months. Among the 183 patients younger than 40 years, only one had pain at 3 months. The authors of the second largest prospective study of zoster,² involving 206 patients from multiple specialty services in Philadelphia, Pennsylvania, found that 14% of patients older than 50 years had pain at 3 months, and 7% had pain at 6 months. In this study, no patients younger than 50 years had pain at 3 months. In comparison, randomized controlled trials of zoster treatment have reported ranges of 16.7% to 60% of patients receiving placebo having pain at 3 months and 5% to 39.1% having pain at 6 months.

Several systematic reviews have addressed treatment or prevention of PHN.^3-5 New trial data^6-9 and discrepancies in the literature^4,10,11 prompted our paper. Differences in the definition of PHN complicated our review of the medical literature.^12,13 For this review, PHN is defined as any pain after cutaneous healing of zoster, and zoster-associated pain (ZAP) is defined as any pain associated with zoster (acute zoster pain and PHN). This review is limited to randomized controlled trials performed at the time of zoster with follow-up addressing the incidence or duration of PHN or ZAP.

Methods

MEDLINE (1966 to present) was searched on December 29, 1998, using PubMed and combining the terms “zoster” or “post-herpetic neuralgia” or “postherpetic neuralgia” or “post-herpetic pain” or “postherpetic pain,” and publication type “clinical trial” (including phases I-IV) or “controlled clinical trial” or “randomized controlled trial.” We searched The Cochrane Controlled Trials Register 1998, issue 4, using the same terms. We also identified trials through article reference lists and Web-based searches. One author conducted independent searches and was unable to identify any additional randomized controlled trials.

For our review we included randomized controlled trials published in English that enrolled primarily immunocompetent subjects with acute herpes zoster and addressed relevant end points (incidence of pain at any time after rash healing in zoster patients or duration of ZAP or PHN). The two authors independently evaluated the trials meeting these inclusion criteria for quality of randomization, allocation concealment, blinding, baseline difference assessment, methods of data collection, adequacy of follow-up (duration and methods), accounting for dropouts, and intention-to-treat analysis. We rated methodologic quality as good, fair, or poor on the basis of an overall assessment of these features. We did not use explicit validity checklists with summary scores, because they have not been shown to predict the effect of bias on treatment differences (ie, they have not been shown to provide more reliable assessments of validity).^14,15 Discordant ratings of trial quality were resolved through the consensus of both investigators.

Results

Our literature search identified 74 appropriate trials. Of the 74 trials reviewed, 42 pertained to treatment at the time of acute zoster to prevent PHN, and 32 pertained to treatment of established PHN.* Of the 42 prevention trials reviewed, 6 were rated as good quality, 22 as fair, and 14 as poor.

Acyclovir

Nucleoside analogues represent the mainstay of acute herpes zoster treatment and produce a faster rate of cutaneous healing and decreased risk of ophthalmic complications.^16,17 A limited subset of acute zoster treatment trials have reported data on PHN incidence or duration.

Four placebo-controlled trials have been conducted using oral acyclovir 800 mg 5 times daily for 7 to 10 days within 72 hours of zoster rash (Table 1).

The largest placebo-controlled trial of acyclovir involved 376 patients from 3 United Kingdom centers and has been reported in multiple publications.^16,18-22 Acyclovir had no effect on the incidence or severity of PHN. Follow-up was monthly to 6 months or until there was no pain during the previous month; this loss to follow-up after the first cessation of pain is the major criticism of this trial, because PHN is known to relapse.^22,23 With this loss to follow-up, the amount of time to complete cessation of pain could not be determined.^23,24 Ten-year follow-up of 132 patients from this trial who reported resolution of pain found that pain returned in 16 cases (12%).²² In subsequent publications, investigators have reported conflicting results in terms of the long-term incidence of PHN. A statistically significant difference in the incidence of PHN was reported in 57 patients from Southampton contacted at 5 years (7% acyclovir patients vs 37% placebo patients, P = .01),²¹ but not in 160 patients from the Sheffield and Birmingham centers contacted at 9 to 10 years.²²

A smaller US multicenter placebo-controlled trial of oral acyclovir reported decreased pain at 1 to 3 months posttreatment, but no differences in analgesic use were reported.^25,26 A reanalysis of 166 patients with pain at enrollment found a reduction in the median duration of ZAP from 62 days to 20 days with acyclovir (P = .02).²⁷ However, of the 21 patients with no pain on enrollment, 15 later developed pain and 4 of those patients continued to report pain through the seventh month. It is possible that inclusion of these patients in the analysis would materially change the results.

In a third trial²⁸ consisting of 83 patients presenting to general practitioners, acyclovir was associated with statistically significant less pain to 3 months. Total analgesic use, however, was reduced in the acyclovir group for only the first 4 weeks.

The fourth trial,²⁹ consisting of 46 patients with acute herpes zoster ophthalmicus, acyclovir was associated with a reduction in pain incidence that was statistically significant at 2 months (P = .04), but not at 6 months (P = .07). Pain severity was significantly reduced from 2 through 6 months, although this analysis is based on only 6 patients having pain at 6 months.

These 4 trials provided all or most of the data for several reviews and meta-analyses.^4,23,24,30 Summary results among these reviews varied because of different methods used, different subgroups assessed, and different efficacy end points reported but generally supported short-term benefit for oral acyclovir. A key issue affecting the overall results of these reviews was consideration of time to complete cessation of pain. This end point was clearly recorded in the 3 smaller trials with positive results but not reported in the large trial with negative results as discussed above.

Crooks and colleagues²³ performed a pooled analysis and reported that oral acyclovir reduced the incidence of PHN from 19% to 11% at 3 to 6 months. We (as well as Lancaster and colleagues⁴) were unable to verify the placebo group incidence rates reported by Crooks and coworkers which appear to have overestimated any acyclovir benefit, and therefore cannot substantiate these findings. Analysis for time to first cessation of pain was reported as a nonsignificant trend toward benefit with acyclovir, while analysis for time to complete cessation of pain (which excluded the largest trial with the negative results) was reported as a statistically significant reduction from an average 86 to 49 days.

Lancaster and colleagues⁴ found a statistically significant reduction in pain at 3 months but no statistically significant benefit at 1 month or 6 months. They commented that analgesic use was not significantly different in 2 of the positive studies.

Wood and coworkers²⁴ derived results from previously unpublished data from these 4 trials and reported that acyclovir significantly accelerated pain resolution. Significant reductions in incidence of pain with acyclovir were reported at 3 and 6 months, but it is unclear if the largest of the 4 trials was included in this analysis.

In the most recent meta-analysis,³⁰ Jackson and colleagues included a fifth trial³¹ and reported an absolute risk reduction of 16% for the incidence of any pain at 6 months (P <.05; number needed to treat [NNT] = 6.3). Multiple methodologic flaws in that meta-analysis, detailed in the Cochrane Review Database,³² make application of the results unclear.

Additional oral acyclovir trials are summarized in Table 1.* Extending the duration of treatment to 14 or 21 days did not provide benefit over 7 days.^33,34 Lower doses of acyclovir had no effect on incidence, severity, or duration of PHN.^17,35,36 There is no evidence to support use of intravenous or topical acyclovir for prevention of PHN.*

Marginal evidence exists to suggest that oral acyclovir 800 mg 5 times daily may reduce the incidence of pain at 1 to 3 months. Information regarding effects on quality of life were generally lacking.

Newer Antivirals

Famciclovir, a prodrug of penciclovir, did not affect the incidence of PHN measured at the time of cutaneous healing in a placebo-controlled trial of 419 patients6 (Table 2). Among the 186 patients (44%) who developed PHN, famciclovir significantly reduced its duration by a median of 2 months, 3.5 months in patients older than 50 years. No dose-response difference was noted. In a subsequent publication presenting monthly prevalence data,³⁷ pain 6 months after enrollment was reported in 15% of patients assigned famciclovir 500 mg and 23.8% of the patients assigned placebo (NNT = 11.4). Similar results were stated for the 750-mg group. Results were reported as statistically significant but P values were not given.

Valacyclovir, a prodrug of acyclovir, has not been studied in a placebo-controlled trial in patients older than 50 years, based on published reports. In a comparison trial with acyclovir,⁷ valacyclovir for 7 or 14 days did not reduce the incidence of PHN (pain after rash healing) but did accelerate its resolution by 1 to 2 weeks. Pain persisting for 6 months was found in 25.7% of the acyclovir group and 18.6% of the combined valacyclovir group (P = .02; NNT = 15.6). The actual benefit of valacyclovir cannot be determined, since the evidence for acyclovir is inconclusive.

Steroids

Inflammation of peripheral nervous system structures has been identified in specimens from patients with PHN.³⁸ Because of this, corticosteroids have been used in the treatment of herpes zoster in hopes of preventing PHN. Trials evaluating steroids are heterogeneous, involving different drugs, doses, routes, durations, and follow-up methods, thus hampering pooling of trial data (Table 3).

Esmann and coworkers³⁹ compared a 21-day oral prednisolone treatment with placebo in 84 patients, all of whom received acyclovir for the first 7 days. Pretrial calculation demonstrated that 324 patients would be required to detect an 80% change in incidence of PHN at 6 months with statistical significance. Enrollment was stopped early when an interim analysis did not show any treatment effect at 3 months at the P <.10 level. No benefit was seen at 6 months. Data points before 6 months were not reported. This is the methodologically strongest of the 4 early steroid trials.

Clemmensen and Andersen⁴⁰ randomized 60 patients to adrenocorticotropic hormone, prednisone, or placebo. No benefit was found for prednisone, and both active treatments may have increased pain at 1 month. No data were reported beyond 6 weeks.

Keczkes and Basheer⁴¹ randomized 40 patients with “severe, painful” zoster to prednisolone or carbamazepine for 4 weeks. A decreased incidence in pain was reported with prednisolone at 8 weeks, but statistics were not reported. Of the 4 early steroid trials, this shows the greatest benefit but is methodologically the weakest study. Without a placebo group, the purported benefit of prednisolone due to a detrimental effect of carbamazepine cannot be excluded.

Eaglstein and colleagues⁴² randomized 35 patients with “severely painful” zoster to triamcinolone versus placebo. No patients younger than 60 years developed neuralgia. Pain reduction with triamcinolone was reported at 8 weeks but not at 6 months for the 24 patients who developed neuralgia. Statistics were not reported.

These 4 trials have undergone several reviews. In a meta-analysis, Lycka10 reported statistically significant reductions in pain at 6 and 12 weeks after zoster onset but not at 6 months. That meta-analysis incorporated unpublished data obtained from the original investigators. NNTs of 4.5 to prevent pain at 6 weeks and 3.5 to prevent pain at 12 weeks were reported, although an intention-to-treat analysis was not performed. The authors of 2 other reviews^4,11 concluded that there was insufficient evidence to support a benefit from steroids, finding the trials too heterogeneous to appropriately combine in meta-analysis.⁴

A later placebo-controlled trial³¹ randomized 201 patients to acyclovir, prednisone, both, or neither for 21 days. That trial had good methodology and may have been large enough to provide the evidence to support or refute the benefit of steroids. No differences in pain incidence were found at 3 or 6 months. Benefit during the first month was found, but results were reported only as risk reductions.With no incidence or duration data reported, these results are not clinically applicable. Follow-up commentary⁴³ pointed out the need for actual data instead of risk ratios. The authors’ response provided limited data showing reductions in median time to cessation of analgesic use (from 28 to 14 days), return to normal activity (from 21 to 3 days), and uninterrupted sleep (from 26 to 5 days) for the combined acyclovir and prednisone treatment group compared with the placebo group. No data were provided for the acyclovir-only or prednisone-only groups.

Wood and colleagues³⁴ randomized 400 patients to prednisolone or placebo for 21 days. Patients were also randomized to acyclovir for 7 days or 21 days, so there was no placebo-only group. There was no benefit to extending acyclovir to 21 days or to adding prednisolone with respect to incidence or duration of PHN.

A recent review⁴⁴ incorporating these newer trials concluded that steroids can reduce acute herpes zoster pain and improve short-term (1-month) quality of life and suggested that steroids are reasonable to use in patients older than 50 years. No effect, however, was demonstrated on the incidence, severity, or duration of PHN.

Tricyclic Antidepressants

Amitriptyline exerts a pain-modulating effect separate from its antidepressant properties⁴⁵ and has been widely used in neuropathic and other chronic pain states. Amitriptyline (25 mg nightly) used preemptively for 90 days starting within 48 hours of rash onset showed a statistically significant reduction in pain incidence at 6 months in a single placebo-controlled trial of fair quality.⁸ Blinding may have been inadequate, because patients were warned of potential dry mouth. Follow-up was a single contact by telephone or mail at 6 to 8 months. Acyclovir given by general practitioners was not controlled. Twenty-four percent of the amitriptyline group and 50% of the placebo group received acyclovir. Among amitriptyline-treated patients, there was a nonsignificant trend toward reduced pain with acyclovir use, while placebo patients who received acyclovir experienced a nonsignificant trend toward increased pain.

Percutaneous Electrical Nerve Stimulation

Percutaneous electrical nerve stimulation (PENS) compared favorably with famciclovir in a single blind randomized trial of 50 adults with zoster.⁴⁶ A 12% absolute risk reduction in pain was reported at 3 and 6 months (but not at 9 months), but no statistical analysis was reported for this measure. Statistically significant reductions in severity of pain were found at 3 and 6 months. Costs and availability of PENS were not reported.

Discussion

There are limited data from randomized controlled trials that indicate that early treatment of acute herpes zoster decreases the incidence or duration of PHN. Acyclovir was the most-studied agent, but there is no convincing evidence that acyclovir alters the course of PHN. There is some evidence that oral acyclovir 800 mg 5 times daily for 7 to 10 days reduces the incidence of pain in the short term (1 to 3 months). Valacyclovir was somewhat more effective than acyclovir in the single largest antiviral trial, but without a placebo control, the actual efficacy of either drug is indeterminate. Famciclovir did not alter the incidence of PHN (at the time of rash healing) but did significantly reduce the duration of PHN in a single placebo-controlled trial of good methodology.

Randomized antiviral trials have been limited totrials including patients presenting within 72 hours of rash onset. No data are available to address the use of antivirals initiated more than 72 hours after rash onset. In the largest prospective study of patients presenting to general practitioners with zoster,¹ only 44% presented within 72 hours of rash onset.

For immunocompetent subjects, oral acyclovir, famciclovir, and valacyclovir were free of major toxicities and demonstrated side effects comparable with placebo in the clinical trials presented. According to the package inserts, posttrial case reports of more serious reactions, including anaphylaxis and renal failure, have been cited. These medications require dose adjustment in patients with compromised renal function.

Steroids, like antivirals, are widely prescribed for the treatment of acute herpes zoster. Although useful for reduction of early pain, there is no evidence that systemic steroids prevent or shorten the course of PHN. No further trials are needed in this regard. The largest and best designed trial involving corticosteroids for the treatment of zoster is rendered clinically inapplicable by the presentation of relative risk reduction rather than incidence and duration data for placebo and treatment groups.

Preemptive amitriptyline and PENS appear promising on the basis of single trials of fair quality. Potentially promising agents based on poor quality trials* include amantadine and Clinacanthus nutans cream. On the basis of the quality of available evidence, availability of therapy, and costs (Table 5), amitriptyline is the most promising of these agents. Anticholinergic side effects are likely to be of greatest significance in the population at risk for PHN (the elderly), although only 3 of 41 patients (7%) withdrew from therapy in the amitriptyline trial.⁸

Limitations

A systematic review of PHN is hampered by different definitions of PHN (ranging from pain immediately following rash healing to 6 months after rash onset), differences in primary end points measured, and differences in study follow-up methods and duration. Although we stated our definition of PHN (pain following rash healing), we have not restricted our analysis to studies employing the same definition or attempted to analyze only study data using that definition. There is debate about whether a pain continuum (ZAP) or subdivision (acute, subacute, and chronic) of zoster-related pain is most suitable for randomized controlled trials examining the impact of treatments on PHN.^47,48 Although the pain that often heralds and more frequently accompanies zoster may merit treatment, we are particularly interested in whether such treatment will have an impact on the pain incidence, duration, and severity that follow rash healing.

We used a comprehensive literature-based search. Searches for unpublished literature and contact with investigators were generally not undertaken. There is a strong chance of publication bias because of our methods of searching. For example, the product information for valacyclovir notes a placebo-controlled trial in patients younger than 50 years presenting with 72 hours of zoster that found no difference with respect to the duration of pain after rash healing. We were unable to locate this trial in a published format, so it did not meet our inclusion criteria.

Language bias is another source of publication basis. We only included studies published in English. Sixteen non–English-language studies identified were potentially pertinent to the prevention of PHN.*

Quality-of-life measures may be ultimately more important than measures of incidence or duration for assessing the impact of treatment on patients. Few trials^7,31,34,46 addressed quality-of-life measures, and these were generally reported only for the short term. Even in the newer antirviral trials demonstrating reduction in the duration of PHN,^6,7 a significant impact on quality of life was not well documented.

Analysis of the power of trials with negative results was not performed, so potential benefits of treatments not studied adequately cannot be excluded.

The expected outcome without treatment is an important consideration in decisions regarding prevention of PHN. There is an inherent selection bias in randomized controlled trials because an unknown number of patients with zoster will not present for medical attention. There is also a reporting bias with identification of PHN in subjects who would otherwise not be troubled enough by symptoms to present for medical care. Therefore, natural history data derived from placebo cohorts of randomized controlled trials is likely to overestimate the true incidence of clinically significant PHN.

Recommendations for future research

Trials of patients with acute herpes zoster should include all the standard criteria for good methodology (adequate randomization, blinding, and so forth), adequate numbers to detect significant differences determined a priori, continued follow-up of all randomized patients for at least 6 months, and detailed descriptions of the studied population. Trials should be limited to subjects older than 50 years, since this is the age group most likely to be afflicted with PHN. Trials should also evaluate patient-oriented end points other than pain, such as quality of life and time to return to usual activity, and should consider analgesic use as a surrogate pain measure. Cost, compliance, and tolerability should be assessed. For clinical applications, P values and NNT should be reported.

A comparison trial of famciclovir and valacyclovir is recommended. Future use of placebo arms in zoster trials is a matter of continued debate,^34,49 especially for trials enrolling patients with ophthalmic zoster.

Future trials of amitriptyline should extend enrollment to 96 hours or more. Inclusion in the amitriptyline trial was limited to subjects presenting within 48 hours of rash onset. This limitation would be very restrictive in clinical practice.

Further investigation of PENS and possibly Clinacanthus nutans cream and amantadine should be considered.

Primary prevention of PHN may best be achieved through prevention of varicella and subsequent zoster through vaccinations. It has been demonstrated in 2 immunized populations (children with leukemia and renal transplant patients) that zoster occurs 5 to 7 times less frequently after vaccination than after natural varicella.⁵⁰ Follow-up of immunocompromised and immunocompetent vaccine recipients is warranted to evaluate the vaccine’s impact on the natural history of zoster and PHN. Waning cell-mediated immunity to varicella-zoster virus with age has been associated with zoster, and booster vaccinations in the elderly have been shown to improve laboratory markers of cell-mediated immunity. Thus, there is a possibility that vaccinations of elderly patients may reduce the impact of zoster and subsequent PHN.⁵¹ Such a trial is currently recruiting 37,000 volunteers older than 60 years who have had chickenpox but have never had shingles.⁵² It should be noted that theoretical arguments have been made that primary vaccination of varicella could lead to subsequent increases in zoster incidence.⁵³

Recommendations for clinical practice

When acute herpes zoster occurs, providers should restrict their attention regarding PHN prevention to patients older than 50 years. According to the available evidence, antiviral therapy (famciclovir or valacyclovir for 7 days) started within 72 hours of rash onset and/or low-dose amitriptyline (for 90 days) may be offered in an effort to reduce PHN incidence or duration. Patients should be informed of the natural history of zoster and PHN, expected benefits for medications, potential for adverse effects, and expected costs. Other treatments, such as steroids or analgesics, may be offered to alter the acute course of zoster but would have no effect on preventing PHN.

Acknowledgments

The authors wish to thank Dehorah Lovett, Karen J. Alper, and Stacey Raulzhan for their assistance.

Postherpetic neuralgia (PHN) is the most common complication of zoster and is much more prevalent among older patients.^1,2 The results of the largest English-language prospective study of patients presenting with zoster,¹ involving 457 patients from 62 general practitioners in Iceland, suggest that the average family physician caring for 2000 patients would expect to see 4 cases of zoster per year and one case of PHN lasting more than 3 months once in 3 years. Among patients older than 60 years in this study, 19% had pain at 3 months, and 8% had pain at 12 months. Among the 183 patients younger than 40 years, only one had pain at 3 months. The authors of the second largest prospective study of zoster,² involving 206 patients from multiple specialty services in Philadelphia, Pennsylvania, found that 14% of patients older than 50 years had pain at 3 months, and 7% had pain at 6 months. In this study, no patients younger than 50 years had pain at 3 months. In comparison, randomized controlled trials of zoster treatment have reported ranges of 16.7% to 60% of patients receiving placebo having pain at 3 months and 5% to 39.1% having pain at 6 months.

Several systematic reviews have addressed treatment or prevention of PHN.^3-5 New trial data^6-9 and discrepancies in the literature^4,10,11 prompted our paper. Differences in the definition of PHN complicated our review of the medical literature.^12,13 For this review, PHN is defined as any pain after cutaneous healing of zoster, and zoster-associated pain (ZAP) is defined as any pain associated with zoster (acute zoster pain and PHN). This review is limited to randomized controlled trials performed at the time of zoster with follow-up addressing the incidence or duration of PHN or ZAP.

Methods

MEDLINE (1966 to present) was searched on December 29, 1998, using PubMed and combining the terms “zoster” or “post-herpetic neuralgia” or “postherpetic neuralgia” or “post-herpetic pain” or “postherpetic pain,” and publication type “clinical trial” (including phases I-IV) or “controlled clinical trial” or “randomized controlled trial.” We searched The Cochrane Controlled Trials Register 1998, issue 4, using the same terms. We also identified trials through article reference lists and Web-based searches. One author conducted independent searches and was unable to identify any additional randomized controlled trials.

For our review we included randomized controlled trials published in English that enrolled primarily immunocompetent subjects with acute herpes zoster and addressed relevant end points (incidence of pain at any time after rash healing in zoster patients or duration of ZAP or PHN). The two authors independently evaluated the trials meeting these inclusion criteria for quality of randomization, allocation concealment, blinding, baseline difference assessment, methods of data collection, adequacy of follow-up (duration and methods), accounting for dropouts, and intention-to-treat analysis. We rated methodologic quality as good, fair, or poor on the basis of an overall assessment of these features. We did not use explicit validity checklists with summary scores, because they have not been shown to predict the effect of bias on treatment differences (ie, they have not been shown to provide more reliable assessments of validity).^14,15 Discordant ratings of trial quality were resolved through the consensus of both investigators.

Results

Our literature search identified 74 appropriate trials. Of the 74 trials reviewed, 42 pertained to treatment at the time of acute zoster to prevent PHN, and 32 pertained to treatment of established PHN.* Of the 42 prevention trials reviewed, 6 were rated as good quality, 22 as fair, and 14 as poor.

Acyclovir

Nucleoside analogues represent the mainstay of acute herpes zoster treatment and produce a faster rate of cutaneous healing and decreased risk of ophthalmic complications.^16,17 A limited subset of acute zoster treatment trials have reported data on PHN incidence or duration.

Four placebo-controlled trials have been conducted using oral acyclovir 800 mg 5 times daily for 7 to 10 days within 72 hours of zoster rash (Table 1).

The largest placebo-controlled trial of acyclovir involved 376 patients from 3 United Kingdom centers and has been reported in multiple publications.^16,18-22 Acyclovir had no effect on the incidence or severity of PHN. Follow-up was monthly to 6 months or until there was no pain during the previous month; this loss to follow-up after the first cessation of pain is the major criticism of this trial, because PHN is known to relapse.^22,23 With this loss to follow-up, the amount of time to complete cessation of pain could not be determined.^23,24 Ten-year follow-up of 132 patients from this trial who reported resolution of pain found that pain returned in 16 cases (12%).²² In subsequent publications, investigators have reported conflicting results in terms of the long-term incidence of PHN. A statistically significant difference in the incidence of PHN was reported in 57 patients from Southampton contacted at 5 years (7% acyclovir patients vs 37% placebo patients, P = .01),²¹ but not in 160 patients from the Sheffield and Birmingham centers contacted at 9 to 10 years.²²

A smaller US multicenter placebo-controlled trial of oral acyclovir reported decreased pain at 1 to 3 months posttreatment, but no differences in analgesic use were reported.^25,26 A reanalysis of 166 patients with pain at enrollment found a reduction in the median duration of ZAP from 62 days to 20 days with acyclovir (P = .02).²⁷ However, of the 21 patients with no pain on enrollment, 15 later developed pain and 4 of those patients continued to report pain through the seventh month. It is possible that inclusion of these patients in the analysis would materially change the results.

In a third trial²⁸ consisting of 83 patients presenting to general practitioners, acyclovir was associated with statistically significant less pain to 3 months. Total analgesic use, however, was reduced in the acyclovir group for only the first 4 weeks.

The fourth trial,²⁹ consisting of 46 patients with acute herpes zoster ophthalmicus, acyclovir was associated with a reduction in pain incidence that was statistically significant at 2 months (P = .04), but not at 6 months (P = .07). Pain severity was significantly reduced from 2 through 6 months, although this analysis is based on only 6 patients having pain at 6 months.

These 4 trials provided all or most of the data for several reviews and meta-analyses.^4,23,24,30 Summary results among these reviews varied because of different methods used, different subgroups assessed, and different efficacy end points reported but generally supported short-term benefit for oral acyclovir. A key issue affecting the overall results of these reviews was consideration of time to complete cessation of pain. This end point was clearly recorded in the 3 smaller trials with positive results but not reported in the large trial with negative results as discussed above.

Crooks and colleagues²³ performed a pooled analysis and reported that oral acyclovir reduced the incidence of PHN from 19% to 11% at 3 to 6 months. We (as well as Lancaster and colleagues⁴) were unable to verify the placebo group incidence rates reported by Crooks and coworkers which appear to have overestimated any acyclovir benefit, and therefore cannot substantiate these findings. Analysis for time to first cessation of pain was reported as a nonsignificant trend toward benefit with acyclovir, while analysis for time to complete cessation of pain (which excluded the largest trial with the negative results) was reported as a statistically significant reduction from an average 86 to 49 days.

Lancaster and colleagues⁴ found a statistically significant reduction in pain at 3 months but no statistically significant benefit at 1 month or 6 months. They commented that analgesic use was not significantly different in 2 of the positive studies.

Wood and coworkers²⁴ derived results from previously unpublished data from these 4 trials and reported that acyclovir significantly accelerated pain resolution. Significant reductions in incidence of pain with acyclovir were reported at 3 and 6 months, but it is unclear if the largest of the 4 trials was included in this analysis.

In the most recent meta-analysis,³⁰ Jackson and colleagues included a fifth trial³¹ and reported an absolute risk reduction of 16% for the incidence of any pain at 6 months (P <.05; number needed to treat [NNT] = 6.3). Multiple methodologic flaws in that meta-analysis, detailed in the Cochrane Review Database,³² make application of the results unclear.

Additional oral acyclovir trials are summarized in Table 1.* Extending the duration of treatment to 14 or 21 days did not provide benefit over 7 days.^33,34 Lower doses of acyclovir had no effect on incidence, severity, or duration of PHN.^17,35,36 There is no evidence to support use of intravenous or topical acyclovir for prevention of PHN.*

Marginal evidence exists to suggest that oral acyclovir 800 mg 5 times daily may reduce the incidence of pain at 1 to 3 months. Information regarding effects on quality of life were generally lacking.

Newer Antivirals

Famciclovir, a prodrug of penciclovir, did not affect the incidence of PHN measured at the time of cutaneous healing in a placebo-controlled trial of 419 patients6 (Table 2). Among the 186 patients (44%) who developed PHN, famciclovir significantly reduced its duration by a median of 2 months, 3.5 months in patients older than 50 years. No dose-response difference was noted. In a subsequent publication presenting monthly prevalence data,³⁷ pain 6 months after enrollment was reported in 15% of patients assigned famciclovir 500 mg and 23.8% of the patients assigned placebo (NNT = 11.4). Similar results were stated for the 750-mg group. Results were reported as statistically significant but P values were not given.

Valacyclovir, a prodrug of acyclovir, has not been studied in a placebo-controlled trial in patients older than 50 years, based on published reports. In a comparison trial with acyclovir,⁷ valacyclovir for 7 or 14 days did not reduce the incidence of PHN (pain after rash healing) but did accelerate its resolution by 1 to 2 weeks. Pain persisting for 6 months was found in 25.7% of the acyclovir group and 18.6% of the combined valacyclovir group (P = .02; NNT = 15.6). The actual benefit of valacyclovir cannot be determined, since the evidence for acyclovir is inconclusive.

Steroids

Inflammation of peripheral nervous system structures has been identified in specimens from patients with PHN.³⁸ Because of this, corticosteroids have been used in the treatment of herpes zoster in hopes of preventing PHN. Trials evaluating steroids are heterogeneous, involving different drugs, doses, routes, durations, and follow-up methods, thus hampering pooling of trial data (Table 3).

Esmann and coworkers³⁹ compared a 21-day oral prednisolone treatment with placebo in 84 patients, all of whom received acyclovir for the first 7 days. Pretrial calculation demonstrated that 324 patients would be required to detect an 80% change in incidence of PHN at 6 months with statistical significance. Enrollment was stopped early when an interim analysis did not show any treatment effect at 3 months at the P <.10 level. No benefit was seen at 6 months. Data points before 6 months were not reported. This is the methodologically strongest of the 4 early steroid trials.

Clemmensen and Andersen⁴⁰ randomized 60 patients to adrenocorticotropic hormone, prednisone, or placebo. No benefit was found for prednisone, and both active treatments may have increased pain at 1 month. No data were reported beyond 6 weeks.

Keczkes and Basheer⁴¹ randomized 40 patients with “severe, painful” zoster to prednisolone or carbamazepine for 4 weeks. A decreased incidence in pain was reported with prednisolone at 8 weeks, but statistics were not reported. Of the 4 early steroid trials, this shows the greatest benefit but is methodologically the weakest study. Without a placebo group, the purported benefit of prednisolone due to a detrimental effect of carbamazepine cannot be excluded.

Eaglstein and colleagues⁴² randomized 35 patients with “severely painful” zoster to triamcinolone versus placebo. No patients younger than 60 years developed neuralgia. Pain reduction with triamcinolone was reported at 8 weeks but not at 6 months for the 24 patients who developed neuralgia. Statistics were not reported.

These 4 trials have undergone several reviews. In a meta-analysis, Lycka10 reported statistically significant reductions in pain at 6 and 12 weeks after zoster onset but not at 6 months. That meta-analysis incorporated unpublished data obtained from the original investigators. NNTs of 4.5 to prevent pain at 6 weeks and 3.5 to prevent pain at 12 weeks were reported, although an intention-to-treat analysis was not performed. The authors of 2 other reviews^4,11 concluded that there was insufficient evidence to support a benefit from steroids, finding the trials too heterogeneous to appropriately combine in meta-analysis.⁴

A later placebo-controlled trial³¹ randomized 201 patients to acyclovir, prednisone, both, or neither for 21 days. That trial had good methodology and may have been large enough to provide the evidence to support or refute the benefit of steroids. No differences in pain incidence were found at 3 or 6 months. Benefit during the first month was found, but results were reported only as risk reductions.With no incidence or duration data reported, these results are not clinically applicable. Follow-up commentary⁴³ pointed out the need for actual data instead of risk ratios. The authors’ response provided limited data showing reductions in median time to cessation of analgesic use (from 28 to 14 days), return to normal activity (from 21 to 3 days), and uninterrupted sleep (from 26 to 5 days) for the combined acyclovir and prednisone treatment group compared with the placebo group. No data were provided for the acyclovir-only or prednisone-only groups.

Wood and colleagues³⁴ randomized 400 patients to prednisolone or placebo for 21 days. Patients were also randomized to acyclovir for 7 days or 21 days, so there was no placebo-only group. There was no benefit to extending acyclovir to 21 days or to adding prednisolone with respect to incidence or duration of PHN.

A recent review⁴⁴ incorporating these newer trials concluded that steroids can reduce acute herpes zoster pain and improve short-term (1-month) quality of life and suggested that steroids are reasonable to use in patients older than 50 years. No effect, however, was demonstrated on the incidence, severity, or duration of PHN.

Tricyclic Antidepressants

Amitriptyline exerts a pain-modulating effect separate from its antidepressant properties⁴⁵ and has been widely used in neuropathic and other chronic pain states. Amitriptyline (25 mg nightly) used preemptively for 90 days starting within 48 hours of rash onset showed a statistically significant reduction in pain incidence at 6 months in a single placebo-controlled trial of fair quality.⁸ Blinding may have been inadequate, because patients were warned of potential dry mouth. Follow-up was a single contact by telephone or mail at 6 to 8 months. Acyclovir given by general practitioners was not controlled. Twenty-four percent of the amitriptyline group and 50% of the placebo group received acyclovir. Among amitriptyline-treated patients, there was a nonsignificant trend toward reduced pain with acyclovir use, while placebo patients who received acyclovir experienced a nonsignificant trend toward increased pain.

Percutaneous Electrical Nerve Stimulation

Percutaneous electrical nerve stimulation (PENS) compared favorably with famciclovir in a single blind randomized trial of 50 adults with zoster.⁴⁶ A 12% absolute risk reduction in pain was reported at 3 and 6 months (but not at 9 months), but no statistical analysis was reported for this measure. Statistically significant reductions in severity of pain were found at 3 and 6 months. Costs and availability of PENS were not reported.

Discussion

There are limited data from randomized controlled trials that indicate that early treatment of acute herpes zoster decreases the incidence or duration of PHN. Acyclovir was the most-studied agent, but there is no convincing evidence that acyclovir alters the course of PHN. There is some evidence that oral acyclovir 800 mg 5 times daily for 7 to 10 days reduces the incidence of pain in the short term (1 to 3 months). Valacyclovir was somewhat more effective than acyclovir in the single largest antiviral trial, but without a placebo control, the actual efficacy of either drug is indeterminate. Famciclovir did not alter the incidence of PHN (at the time of rash healing) but did significantly reduce the duration of PHN in a single placebo-controlled trial of good methodology.

Randomized antiviral trials have been limited totrials including patients presenting within 72 hours of rash onset. No data are available to address the use of antivirals initiated more than 72 hours after rash onset. In the largest prospective study of patients presenting to general practitioners with zoster,¹ only 44% presented within 72 hours of rash onset.

For immunocompetent subjects, oral acyclovir, famciclovir, and valacyclovir were free of major toxicities and demonstrated side effects comparable with placebo in the clinical trials presented. According to the package inserts, posttrial case reports of more serious reactions, including anaphylaxis and renal failure, have been cited. These medications require dose adjustment in patients with compromised renal function.

Steroids, like antivirals, are widely prescribed for the treatment of acute herpes zoster. Although useful for reduction of early pain, there is no evidence that systemic steroids prevent or shorten the course of PHN. No further trials are needed in this regard. The largest and best designed trial involving corticosteroids for the treatment of zoster is rendered clinically inapplicable by the presentation of relative risk reduction rather than incidence and duration data for placebo and treatment groups.

Preemptive amitriptyline and PENS appear promising on the basis of single trials of fair quality. Potentially promising agents based on poor quality trials* include amantadine and Clinacanthus nutans cream. On the basis of the quality of available evidence, availability of therapy, and costs (Table 5), amitriptyline is the most promising of these agents. Anticholinergic side effects are likely to be of greatest significance in the population at risk for PHN (the elderly), although only 3 of 41 patients (7%) withdrew from therapy in the amitriptyline trial.⁸

Limitations

A systematic review of PHN is hampered by different definitions of PHN (ranging from pain immediately following rash healing to 6 months after rash onset), differences in primary end points measured, and differences in study follow-up methods and duration. Although we stated our definition of PHN (pain following rash healing), we have not restricted our analysis to studies employing the same definition or attempted to analyze only study data using that definition. There is debate about whether a pain continuum (ZAP) or subdivision (acute, subacute, and chronic) of zoster-related pain is most suitable for randomized controlled trials examining the impact of treatments on PHN.^47,48 Although the pain that often heralds and more frequently accompanies zoster may merit treatment, we are particularly interested in whether such treatment will have an impact on the pain incidence, duration, and severity that follow rash healing.

We used a comprehensive literature-based search. Searches for unpublished literature and contact with investigators were generally not undertaken. There is a strong chance of publication bias because of our methods of searching. For example, the product information for valacyclovir notes a placebo-controlled trial in patients younger than 50 years presenting with 72 hours of zoster that found no difference with respect to the duration of pain after rash healing. We were unable to locate this trial in a published format, so it did not meet our inclusion criteria.

Language bias is another source of publication basis. We only included studies published in English. Sixteen non–English-language studies identified were potentially pertinent to the prevention of PHN.*

Quality-of-life measures may be ultimately more important than measures of incidence or duration for assessing the impact of treatment on patients. Few trials^7,31,34,46 addressed quality-of-life measures, and these were generally reported only for the short term. Even in the newer antirviral trials demonstrating reduction in the duration of PHN,^6,7 a significant impact on quality of life was not well documented.

Analysis of the power of trials with negative results was not performed, so potential benefits of treatments not studied adequately cannot be excluded.

The expected outcome without treatment is an important consideration in decisions regarding prevention of PHN. There is an inherent selection bias in randomized controlled trials because an unknown number of patients with zoster will not present for medical attention. There is also a reporting bias with identification of PHN in subjects who would otherwise not be troubled enough by symptoms to present for medical care. Therefore, natural history data derived from placebo cohorts of randomized controlled trials is likely to overestimate the true incidence of clinically significant PHN.

Recommendations for future research

Trials of patients with acute herpes zoster should include all the standard criteria for good methodology (adequate randomization, blinding, and so forth), adequate numbers to detect significant differences determined a priori, continued follow-up of all randomized patients for at least 6 months, and detailed descriptions of the studied population. Trials should be limited to subjects older than 50 years, since this is the age group most likely to be afflicted with PHN. Trials should also evaluate patient-oriented end points other than pain, such as quality of life and time to return to usual activity, and should consider analgesic use as a surrogate pain measure. Cost, compliance, and tolerability should be assessed. For clinical applications, P values and NNT should be reported.

A comparison trial of famciclovir and valacyclovir is recommended. Future use of placebo arms in zoster trials is a matter of continued debate,^34,49 especially for trials enrolling patients with ophthalmic zoster.

Future trials of amitriptyline should extend enrollment to 96 hours or more. Inclusion in the amitriptyline trial was limited to subjects presenting within 48 hours of rash onset. This limitation would be very restrictive in clinical practice.

Further investigation of PENS and possibly Clinacanthus nutans cream and amantadine should be considered.

Primary prevention of PHN may best be achieved through prevention of varicella and subsequent zoster through vaccinations. It has been demonstrated in 2 immunized populations (children with leukemia and renal transplant patients) that zoster occurs 5 to 7 times less frequently after vaccination than after natural varicella.⁵⁰ Follow-up of immunocompromised and immunocompetent vaccine recipients is warranted to evaluate the vaccine’s impact on the natural history of zoster and PHN. Waning cell-mediated immunity to varicella-zoster virus with age has been associated with zoster, and booster vaccinations in the elderly have been shown to improve laboratory markers of cell-mediated immunity. Thus, there is a possibility that vaccinations of elderly patients may reduce the impact of zoster and subsequent PHN.⁵¹ Such a trial is currently recruiting 37,000 volunteers older than 60 years who have had chickenpox but have never had shingles.⁵² It should be noted that theoretical arguments have been made that primary vaccination of varicella could lead to subsequent increases in zoster incidence.⁵³

Recommendations for clinical practice

When acute herpes zoster occurs, providers should restrict their attention regarding PHN prevention to patients older than 50 years. According to the available evidence, antiviral therapy (famciclovir or valacyclovir for 7 days) started within 72 hours of rash onset and/or low-dose amitriptyline (for 90 days) may be offered in an effort to reduce PHN incidence or duration. Patients should be informed of the natural history of zoster and PHN, expected benefits for medications, potential for adverse effects, and expected costs. Other treatments, such as steroids or analgesics, may be offered to alter the acute course of zoster but would have no effect on preventing PHN.

Acknowledgments

The authors wish to thank Dehorah Lovett, Karen J. Alper, and Stacey Raulzhan for their assistance.