Is acupuncture a risk factor for hepatitis? Systematic review of epidemiological studies

Authors


Professor E Ernst, Complementary Medicine, Peninsula Medical School, Universities of Exeter and Plymouth, 25 Victoria Park Road, Exeter EX2 4NT UK. Email: Edzard.Ernst@pms.ac.uk

Abstract

Objective:  Acupuncture has been repeatedly associated with infectious hepatitis. The aim of the present systematic review was therefore to critically evaluate such data from epidemiological investigations.

Methods:  Four independent literature searches were carried out to identify all epidemiological evidence linking acupuncture with hepatitis. All studies were validated by the authors and data extracted according to predefined criteria.

Results:  Fifteen investigations fulfilled our inclusion criteria. Most studies originated from Asia. One study included markers to hepatitis A virus, three to hepatitis B virus, and 13 to hepatitis C virus. Five investigations reported associations between acupuncture and seropositivity to hepatitis C virus. In those studies, acupuncture increased the risk only modestly.

Conclusion:  A modest association between hepatitis C and acupuncture has been reported in some countries. This emphasizes the importance of exclusively using disposable acupuncture needles.

INTRODUCTION

Acupuncture is an ancient Chinese form of treatment typically involving the insertion of fine steel needles into specific points on the body surface. Its use worldwide is generally widespread. Survey data suggest that prevalence figures in the general population range from 1% in the USA to 21% in France.1 In many parts of Asia, its use is probably almost ubiquitous. Although prospective studies have shown acupuncture to be relatively safe,2–4 numerous case reports revealing serious complications, including infection with hepatitis, continue to emerge.5 Small outbreaks of hepatitis infections have been associated with the practice of acupuncture both in industrialized and developing countries.6–11

To more accurately understand the risks of hepatitis infection after acupuncture treatment, a systematic assessment may be helpful. The purpose of the present review was to critically evaluate epidemiological studies examining the relationship between markers of hepatitis infections and previous use of acupuncture.

METHODS

Literature searches were carried out in the following databases: MEDLINE, Embase, Cochrane Library and Amed (all from their respective inception to December 2001). The key words used were: acupuncture, hepatitis, and epidemiology. In addition, our own files and major reference books of acupuncture were searched for relevant articles. The bibliographies of the articles thus found were scanned for further relevant articles.

Epidemiological cohort, case-control, or cross-sectional studies from a defined population that evaluated the relationship between markers of hepatitis A, B or C infections with the previous use of acupuncture were included. Studies were excluded if they lacked an appropriate comparison group or sufficient details about the study populations.12–19 Dual publications were eliminated. No language restrictions were applied. All included articles were read by both authors. Association was measured by odds ratios, relative risk or P-values. Key data were evaluated and extracted according to predefined criteria (Table 1).

Table 1.   Epidemiological investigations of the association between hepatitis infections and acupuncture
First authorType of investigationStudy population (country)Outcome measureMain findingsComment
  1. HBV, hepatitis B virus; HCV, hepatitis C virus; HGV, hepatitis G virus; OR, odds ratio.

Beaseley 19(1983)Prospective cohort704 college students without evidence of prior hepatitis A infection. 738 college students without evidence of prior hepatitis B infection (Taiwan)Percentage who used acupuncture among seroconverters and of non-seroconvertersHepatitis A: None of 32 seroconverters and 1.6% non-seroconverters used acupuncture. Hepatitis B: 2.3 of seroconverters vs 5.4% of non-converters used acupuncturePopulation was young, healthy and few had previously used acupuncture. There was a high likelihood of hepatitis A and B infection early in life
Phoon (1988)20Cross-sectional6328 Chinese men without evidence of liver disease (Singapore)Serum marker for hepatitis B carrier status and past use of acupunctureNo relationship between previous use and hepatitis B carriers (OR = 0.9, 95% CI = 0.7–1.1)Moderate exposure to acupuncture (21%) among seronegative individuals
Kiyosawa (1994)21Two cross-sectional435 inhabitants of an area with endemicity of HCV, 1542 inhabitants of a non-endemic area (Japan)Use of acupuncture and/or suidama (‘folk remedies’) in HCV positive and negative individuals in both areas HCV positive individuals in endemic area are significantly more likely to have a history of folk remedies than those without HCV; no association in non-endemic areaSuidama is a folk remedy involving cutting the skin and cupping over the wound. Did not distinguish between history of acupuncture and suidama; modest use of folk remedies in controls (17–26%)
Neal (1994)22Case-control74 HCV positive blood donors and 150 matched controls (UK)Previous use of acupuncture in both cases and controlsNo association between acupuncture use and HCV positivity (OR = 1.0, 95% CI = 0.1–6.7)Only 3% of blood donor controls had previously used acupuncture
Hou (1995)23Cross-sectional84 hemodialysis patients, 186 family members of those patients (Taiwan)Serum markers for hepatitis C and history of acupunctureNo statistically significant relation between history of acupuncture and seropositivity in either groupModerate history of acupuncture (9% in families; 20% dialysis patients)
Sulaiman (1995)24Cross-sectional7572 healthy blood donors (Indonesia)Serum markers for hepatitis B and C carriers and history of acupuncture treatmentNo statistically significant relationship between HBV or HCV surface antigen and history of acupuncture, although 13% of HCV carriers had used acupuncture vs 5.8% of controlsPrevious use of acupuncture is uncommon in controls for hepatitis B (7%) and for hepatitis C (6%)
Kim (1996)25Case-control64 HCV positive cases and 128 matched controls from one hospital (Korea)Previous use of acupuncture in both cases and controlsNo association between acupuncture use and HCV positivity (OR = 0.7, not statistically significant)Acupuncture use ubiquitous (80%) in controls
Kayaba (1998)26Cross-sectional2231 participants in mass screening examinations (Japan)Use of acupuncture in HCV positive and negative individualsRelative risk for HCV positivity was 1.30 (95% CI = O.7–2.6) in individuals with history of acupuncture use compared to those without Moderate use of acupuncture (24%) among seronegative individuals
Sun (1999)27Nested case-control272 HCV positive and 282 matched HCV negative controls selected from 11904 men participating in a nationwide survey (Taiwan)History of acupuncture use in HCV serapositive and seronegative menAcupuncture use was 2.4 times more common in seropositive men (adjusted 95% CI = 1.4–4.2) and a population attributable risk fraction of 16%Low (8%) use of acupuncture. Other risk factors were blood transfusions, medical injections and tattooing
Fujiwara (2000)28Cross-sectional6121 participants in Adult Health Study of atomic bomb survivors (Japan)Serum markers for hepatitis C and history of acupunctureNo association between acupuncture use and HCV positivity (OR = 1.0; 95% CI = 0.9–1.1)Forty-one percent had history of acupuncture use
Sanchez (2000)29Cross-sectional1438 people including blood donors, drug abusers, hemophiliacs, patients undergoing dialysis, and patients with hepatitis (Peru)Serum markers for hepatitis C and history of acupunctureAcupuncture more common among HCV seropositive patients (OR = 2.1,
P = 0.06)
History of acupuncture not as large a risk as major surgeries, transfusions, or transplants
Shin (2000)30Cross-sectional and a nested case-matched control study1033 volunteers from a rural area with a high incidence of liver cancer; 64 seropositive cases and 252 seronegative controls (Korea)Association of history of acupuncture and seropositivity to HCV in all volunteers and in case-control studyOdds ratio in entire cohort was 3.3 (95% CI = 2.0–5.5) with evidence of dose–response relationship; the odds ratio in matched pair comparison was 1.2 (95% CI 0.6–2.2) for < 10 sessions, and 2.2 (95% CI
1.0–4.7) for > 10 sessions of acupuncture
Authors suggest that HCV infection was spread through infected acupuncture needles
Chen (1995)31Case-control38 consecutive patients (25 men, 13 women) with acute HCV infections and 76 age and sex matched healthy controls (Taiwan)Previous use of acupuncture and serum markers for hepatitis CFive percent of cases vs 1% of controls used acupuncture in the past; there was no significant association between acupuncture and hepatitis infectionInjection with non-disposable needles was the only significant risk factor (OR = 4.2, 95% CI 1.2–14.5)
Mori (1998)32Cross-sectional1151 males and 2429 females (Japan)Previous use of acupuncture and serum markers for hepatitis CAge-adjusted OR for anti-HCV seropositivity was higher for those who had used acupuncture. OR males 2.5 (95% CI = 1.3–4.7), OR females 1.8 (95% CI = 1.1–3.0)History of blood transfusion and family history of liver disease were significant risk factors
Haley(2001)33Cross-sectional626 consecutive patients evaluated for spinal problems (USA)Previous use of acupuncture and serum markers for hepatitis C6.9% of seronegative and none of 33 seropositive patients had had acupuncture (P = 0.09)Four independent risk factors were identified: infection from drug use, aucullary hospital jobs held by men, tattoos from commercial parlors, and drinking 3–6 packs of beer/month

RESULTS

Fifteen studies fulfilled our inclusion criteria.19–33 Most investigations originated from Asia, especially Taiwan19,23,27,31 and Japan21,26,28,32 but one study was from the UK22 and one was from Peru.29 One study included markers to hepatitis A virus,19 three included markers to hepatitis B virus19,20,24 and 13 included markers to hepatitis C virus (HCV).21–33 Five of the studies in the latter group found an association between previous acupuncture and seropositivity to hepatitis C and these originated from Japan,21,32 Taiwan,27 Peru29 and Korea.30 No positive associations with hepatitis A or B were observed. One of the positive studies did not distinguish between two different folk remedies, suidama and acupuncture,21 and found a positive association only in the area of high hepatitis C endemicity.

In those studies where an increased risk was found, acupuncture increased the risk only modestly and was not the most common risk factor for hepatitis C infection. One study showed that those individuals who reported receiving more than 10 sessions of acupuncture were more likely to be HCV seropositive than those who reported receiving fewer treatments,30 which suggests that increased exposure to acupuncture increased risk of HCV seropositivity. In populations where acupuncture treatment was uncommon (< 10% of controls reported use), only one of six studies found an association. In populations where acupuncture was used by 10–40% of people, none of the two studies were positive. Of the five studies conducted in populations where more than 40% used acupuncture, three found a positive association.

DISCUSSION

Our findings suggest that acupuncture is only a modest risk factor for hepatitis C infections in those populations where a positive association is found. Even though case reports document a potential relationship between acupuncture and hepatitis B infection,6 epidemiological studies do not corroborate this. Similarly, there is little epidemiological support for a link between acupuncture and hepatitis C.

One would expect that the risk of infection from acupuncture would be eliminated by using disposable needles. This is also indirectly supported by epidemiological evidence: a positive association was found only in non-Western countries where the use of disposable needles may not be fully established.21,27,29,30,32 The evidence from case reports mostly precedes the widespread use of disposable needles6–9 or pertains to a form of acupuncture that employs hypodermic needles for infections into acupuncture points.11

Only one study evaluated risk by the number of acupuncture treatments, but these results indicated a dose-effect relationship.30 Because the other studies did not report the level of exposure to acupuncture, this finding cannot be confirmed. In fact, if a causal relationship exists, one would also expect a dose-effect relationship between acupuncture use and hepatitis risk.

Our analysis has several important weaknesses. First, only 15 investigations were located and only one was prospective. Thus there is uncertainty whether acupuncture treatments actually preceded the infection. Second, reporting of important details (e.g. level of exposure) is often insufficient. Third, positive publication bias exists; epidemiological studies of hepatitis risk that failed to find an association with acupuncture may not mention acupuncture at all. These limitations suggest that acupuncture is even less of a risk factor for hepatitis than we reported.

In conclusion, a modest association between hepatitis C and acupuncture was found in some epidemiological studies from developing countries and Asia, whereas other studies found no evidence of an association. The use of disposable acupuncture needles everywhere should eliminate this risk in all populations.

Ancillary