Acupuncture for treating dry eye: a systematic review

Authors


Myeong Soo Lee, PhD
Department of Medical Research
Korea Institute of Oriental Medicine
Daejeon, 305-811
South Korea
Tel: 82 042 868 9266
Fax: 82 042 863 9464
Email: drmslee@gmail.com; mslee@kiom.re.kr

Abstract.

Purpose:  The objective of this review was to assess the effectiveness of acupuncture as a treatment option for treating the condition of dry eye.

Methods:  We searched the literature using 14 databases from their inceptions to 3 December 2009, without language restrictions. We included randomized clinical trials (RCTs) comparing acupuncture with conventional treatment. Their risk of bias was assessed using Cochrane criteria.

Results:  Six RCTs met all the inclusion criteria. Three RCTs compared the effects of acupuncture with artificial tears in patients with xerophthalmia or Sjögren syndrome. A meta-analysis of these data showed that acupuncture improved tear break-up times (p < 0.0001), Schirmer test scores (p < 0.00001), response rates (p = 0.002) and the region of cornea fluorescent staining (p = 0.0001) significantly more than artificial tears did. The other three RCTs compared the effects of acupuncture plus artificial tears with artificial tears alone. Two of these studies failed to show significant effects of acupuncture, while one reported significant effects. For Schirmer test scores and frequency of artificial tear usage, two RCTs reported superior effects of acupuncture plus artificial tears, while one RCT failed to do so.

Conclusion:  These results provide limited evidence for the effectiveness of acupuncture for treating dry eye. However, the total number of RCTs, the total sample size and the methodological quality were too low to draw firm conclusions.

Introduction

The prevalence of dry eye is 11–17% in the general population, and rates of up to 29% have been reported from clinical optometry practice (Reddy et al. 2004). The conventional treatment is mostly tear replacement by topical artificial tears (Calonge 2001). Patients with eye diseases often turn towards complementary and alternative medicine (CAM). Two surveys reported CAM utilization rates for glaucoma and inflammatory eye disease (Rhee et al. 2002; Smith et al. 2004). One of them showed the use of CAM by 5.4% (54 of 1027) of patients with glaucoma in the USA (Rhee et al. 2002), while the other reported the use of CAM by 42% (37 of 89) of patients with inflammatory eye disease in Australia (Smith et al. 2004). However, there were no reports of CAM utilization rates for eye diseases in other countries. Currently, it is not clear whether ethnicity influences the usage of CAM for eye diseases.

Acupuncture is one of the most popular forms of CAM used. It is now a widely accepted intervention for the treatment of a variety of conditions (Ernst 2006). Several studies have indicated that acupuncture is one option for eye diseases (Calonge 2001; O’Brien 2004; Smith et al. 2004). One reason for using acupuncture as a treatment for eye disease is that it is claimed to be able to modulate the autonomic nervous system and immune system (Kavoussi & Ross 2007; Backer et al. 2008; Uchida & Hotta 2008), which in turn might regulate lacrimal gland function. It therefore seems pertinent to evaluate the effectiveness of acupuncture as a treatment for dry eye. Currently, no systematic review of this subject is available. The objective of this systematic review was to summarize and critically assess the evidence from randomized clinical trials (RCTs) of acupuncture for treating dry eye.

Methods

Data sources

The following databases were searched from their respective inceptions through 3 December 2009: MEDLINE, AMED, CINAHL, EMBASE, PsychInfo, Health Technology Assessments, DARE, five Korean Medical Databases (Korean Studies Information, DBPIA, Korea Institute of Science and Technology Information, KoreaMed, and Research Information Service System), Chinese Medical Databases (China National Knowledge infrastructure: CNKI) and The Cochrane Library 2009, Issue 4. The search terms used were ‘acupuncture AND (dry eye OR xerophthalmia OR keratoconjunctivitis sicca)’. In addition, our own files and journals [Focus on Alternative and Complementary Therapies (FACT) and Research in Complementary Medicine (Forschende Komplementarmedizin) up to December 2009] were manually searched. Hard copies of all articles were obtained and read in full.

Study selection

Prospective RCTs of acupuncture for the treatment of dry eye of any origin experienced by human patients were located. Trials were included if they employed needle acupuncture as the sole treatment or as an adjunct to other treatments (if the control group also received the same concomitant treatment as the acupuncture group). Trials testing other forms of acupuncture, such as laser acupuncture or moxibustion, were excluded. Trials comparing any type of acupuncture with any type of control intervention were included. We also included trials that employed acupuncture as an adjunct to conventional treatment. Trials with acupuncture as a part of a complex intervention were excluded. Trials were excluded if dry eye was not a central symptom of the condition. The control interventions were other active interventions without acupuncture or sham acupuncture or no intervention at all. Studies comparing two different forms of acupuncture and those in which no clinical data were reported were also excluded. Dissertations and abstracts were also included. No language restrictions were imposed.

Data extraction, quality and assessment of risk of bias

All articles were read by three independent reviewers (M.S.L., B.C.S. and T.Y.C) who extracted data from the articles according to predefined criteria. Risk of bias was assessed using the Cochrane classification in four criteria: randomization, blinding, withdrawals and allocation concealment (Higgins & Altman 2008). Considering that it is virtually impossible to blind therapists to the use of acupuncture, we assessed patient and assessor blinding separately. Disagreements were resolved by discussion between the three reviewers (M.S.L., B.C.S. and T.Y.C). There were no disagreements between the three reviews about the results.

Data synthesis

To summarize the effects of acupuncture on response rate, we abstracted the risk estimates (relative risk, RR) and 95% confidence intervals from each study. Weighted mean differences (WMD) and 95% confidence intervals (CIs) were also calculated using the Cochrane Collaboration’s software [Review Manager (RevMan) Version 5.0 for Windows, Copenhagen: The Nordic Cochrane Centre] to summarize the effects of acupuncture on continuous outcome variables (BUT and Schirmer test). The chi-square test, tau2 and the Higgins I2 test were used to assess heterogeneity.

Results

Study description

The searches identified 72 potentially relevant articles of which 66 studies were excluded (Fig. 1). Key data of the six included RCTs are summarized in Table 1 (Nepp et al. 1998; Pang et al. 2003; Gronlund et al. 2004; He et al. 2004; Wang et al. 2005; Tseng et al. 2006). Four trials originated from China (Pang et al. 2003; He et al. 2004; Wang et al. 2005; Tseng et al. 2006), one trial was conducted in Sweden (Gronlund et al. 2004), and one trial was from Austria (Nepp et al. 1998). Three of the included trials followed a two-arms parallel group design (Pang et al. 2003; Gronlund et al. 2004; He et al. 2004), two followed three-arms parallel group design (Wang et al. 2005; Tseng et al. 2006), and one followed a four-arms parallel group design (Nepp et al. 1998). All trials adopted the principles of traditional Chinese Medicine as the rationale for selecting acupuncture. Four RCTs (Nepp et al. 1998; Pang et al. 2003; Gronlund et al. 2004; Tseng et al. 2006) used a fixed selection of acupuncture points for all patients, while two trials (He et al. 2004; Wang et al. 2005) employed individualized acupuncture according to pattern classification. The number of treatment sessions ranged from 10 to 30 (Table 2).

Figure 1.

 Flowchart of the trial selection process. UOS = uncontrolled observational study; CCT = controlled clinical trial; RCT = randomized clinical trial.

Table 1.   Summary of parallel group RCTs of acupuncture for treating dry eye.
First author (year)Sample size
condition
Gender (M/F)
mean
age (range)
duration (year)
Intervention group (regime)Control group (regime)Main outcomesIntergroup differencesRisk of bias*
  1. * Risk of bias (sequence generation, incomplete outcome measures, patient blind, assessor blind, allocation concealment), Y = low risk of bias; U = unclear; N = high risk of bias; n.r. = not reported; NS = not significant; CFS = cornea fluorescent staining; BUT = break-up time of tears; SSP = silver spike point; RBS = Rose-Bengal solution; RCT = randomized clinical trials.

Wang et al. (2005) China45 Xerophthalmia (90 eyes)17/28
51.7 (24–75)
0.17–10
(A) AT (individualized points, one session  = 20–25 min, once every 2 days for 10 times, 10 days rest, total three sessions, n = 15)
(B) AT (standardized points, one session = 20–25 min, once every 2 days for 10 times, total three sessions, n = 15)
(C) Artificial tears (five times daily for 30 days, n = 15)(1) BUT
(2) Schirmer I test
(3) Response rate
(4) CFS
(1) A versus C; p < 0.01, B versus C; NS, in favour of AT
(2) A versus C; p < 0.01, B versus C; p < 0.01, in favour of AT
(3) A: 13:2 (87%); B: 10:5 (67%); C: 5:10 (33%) A versus C; p < 0.01, B versus C; NS
(4) A versus C; p < 0.01, B versus C; p < 0.01, in favour of AT
Y, U, U, U, U
He et al. (2004) China32 Xerophthalmia (64 eyes)12/20
52 (35–74)
0.17–10
AT (20–25 min, one session = once every 2 days for 10 days, 10 days rest, total three sessions, n = 16)Artificial tears (5–6 times daily for 30 days, n = 16)(1) BUT
(2) Schirmer test
(3) CFS
(1) p < 0.05 in favour of AT
(2) p < 0.01 in favour of AT
(3) p < 0.05, in favour of AT
U, U, U, U, U
Pang et al. (2003)China60 Sjögren syndrome
(120 eyes)
6/54
22–65
0.5–15
AT (one session = 30 min, daily for 10 times, total three sessions, n = 30)Artificial tears (six times daily for 30 days, n = 30)(1) BUT
(2) Schirmer test
(3) Response rate
(4) CFS
(1) p < 0.05 in favour of AT
(2) p < 0.05 in favour of AT
(3) 63% versus 33%; p < 0.05
(4) p < 0.05 in favour of AT
Y, U, U, U, U
Tseng et al. (2006)China43 Dry eye syndrome16/27
48/52/51
3.24/2.65/4.0
AT (20 min, two times weekly for 8 weeks, n = 17), plus artificial tearsArtificial tears (n = 9)(1) BUT
(2) Schirmer I test
(3) No applications of artificial tears
(1) NS
(2) p < 0.01, in favour of AT
(3) p < 0.01, in favour of AT
Y, U, U, Y, U
Gronlund et al. (2004) Sweden25 Keratoconjunctivitis sicca5/20
56 (31–79)
n.r.
AT (30 min, once or twice weekly, total 10 sessions, n = 12), plus artificial tearsArtificial tears (n = 13)(1) BUT
(2) Schirmer test
(3) Eye drop use
(4) Total symptoms score
(5) RBS recording
(1–5) NSU, Y, U, U, U
Nepp et al. (1998)Austria114 Dry eye syndrome
(228 eyes)
15/99 57.6
(20–84)
n.r.
AT (30 min, once weekly for 10 weeks, n = 30), plus artificial tearsArtificial tears (n = 22)(1) Response rate (BUT)
(2) Response rate (Schirmer test)
(3) Response rate (No eye drop use)
(1–3) p < 0.01 in favour of ATU, U, U, U, U
Table 2.   Summary of treatment acupuncture points and other information.
First author (year)Acupuncture methodTotal treatment (sessions)Acupuncture pointsDe-qiRationale (s) for acupointsAdverse events
  1. n.r. = not reported.

Wang et al. (2005)(A) Individualized by pattern classification
(B) Standardized local acupoints
(A) 30
(B) 30
(A) Heat ablaze and yin injury: Quchi (LI 11), Hegu (LI 4),Sanyinjiao (SP 6), Taixi (KI 3),Sibai (ST 2), Yingxiang (LI 20) Phlegm and blood stasis: Sibai (ST 2), Xuehai (SP 10), Yinlingquan (SP 9), Zusnali (ST 36),Sanyinjiao (SP 6),Fengrong (ST 40)
(B) Cuanzhu (BL 2), Sizhukong (TE 23), Yangbai (GB 14),Chengqi (ST 1)
n.r.TCM theoryn.r.
He et al. (2004)Individualized by pattern classification30Heat ablaze and yin injury: Sibai (ST 2),Yingxiang (LI 20), Quchi (LI 11), Hegu (LI 4),Sanyinjiao (SP 6) Phlegm and blood stasis: Sibai (ST 2),Xuehai (SP 10), Yinlingquan (SP 9),Zusnali (ST 36),Sanyinjiao (SP 6), Zhaohai (KI 6)n.r.TCM theoryn.r.
Pang et al. (2003)Fixed classic acupuncture30Sibai (ST2), Zanzhu (BL2), Chengqi (ST1), Yinxiang (LI20), Baihui (GV20), Shaoze (SI1), Houxi (SI3)n.r.TCM theoryn.r.
Tseng et al. (2006)Fixed classic acupuncture16Taiyang (Ex-HN), Sizhkong (TE 23), Yangbai (GB 14), Sibai (ST 2),Sanyinjiao (SP 6) (bilateral)YesTCM theoryn.r.
Gronlund et al. (2004)Fixed classic acupuncture10ST2, ST8, ST36, GB1, GB14, BL2, LI4n.r.TCM theoryNone
Nepp et al. (1998)Fixed classic acupuncture10Local points: GB1, BL2, ST5, EX2 (Yintang) Specific points for the eyes and mucosa: LI4, SI3, LI3, KI6, TE5n.r.TCM theory and neurohormonal theoryn.r.

Risk of bias

Three RCTs employed appropriate methods of randomization (Pang et al. 2003; Wang et al. 2005; Tseng et al. 2006). Assessor blinding was judged to have been achieved in one of the RCTs (Tseng et al. 2006), and none used allocation concealment. Sufficient details of dropouts and withdrawals were described in one RCT (Gronlund et al. 2004). Adverse events were mentioned in one study (Gronlund et al. 2004).

Outcomes

Acupuncture versus artificial tears

Three RCTs compared the effects of acupuncture for dry eye symptoms with those of artificial tears in patients with xerophthalmia (He et al. 2004; Wang et al. 2005) or Sjogren syndrome (Pang et al. 2003). The meta-analysis of these data showed significant improvements in tear break-up time (BUT; n = 274, WMD, 1.26, 95% CIs 0.67–1.85, p < 0.0001, Fig. 2A), Schirmer test (n = 274, WMD, 2.36, 95% CIs 1.58–3.15, p < 0.00001, Fig. 2B), response rate (n = 105, RR, 2.04, 95% CIs 1.30–3.21, p = 0.002, Fig. 2C) and the region of cornea fluorescent staining (CFS, n = 274, RR, 2.84, 95% CIs 1.67–4.82, p = 0.0001, Fig. 2D). Subgroup analysis also showed favourable effects of acupuncture on BUT (n = 154, WMD, 1.25; 95% CI 0.48–2.02, p = 0.002, Fig. 2A), Schirmer test (n = 154, WMD, 2.43; 95% CI 1.34–3.52, p < 0.0001, Fig. 2B), response rate (n = 45, RR, 2.29, 95% CIs 1.10–4.78, p = 0.03, Fig. 2C) and CFS (n = 154, RR, 2.72, 95% CI 1.48–4.99, p = 0.001, Fig. 2D) in patients with xerophthalmia (He et al. 2004; Wang et al. 2005).

Figure 2.

 Forest plot of acupuncture for (A) BUT, (B) ST, (C) response rate and (D) CFS of dry eye compared with artificial tears. BUT = tear break-up time; CFS = cornea fluorescent staining; ST = Schirmer test.

Acupuncture plus artificial tears versus artificial tears

Three RCTs (Nepp et al. 1998; Gronlund et al. 2004; Tseng et al. 2006) compared the effects of acupuncture plus artificial tears with artificial tears alone. Two RCTs failed to show superior effects of acupuncture plus artificial tears on BUT (Gronlund et al. 2004; Tseng et al. 2006), while one RCT reported significant effects (Nepp et al. 1998). For the Schirmer test and frequency of artificial tear usage, two RCTs reported superior effects of acupuncture plus artificial tears (Nepp et al. 1998; Tseng et al. 2006), while one RCT failed to do so (Gronlund et al. 2004).

Discussion

Few rigorous trials have tested the effects of acupuncture for dry eye. Their results showed suggestive evidence for the effectiveness of acupuncture on dry eye compared with artificial tears (Pang et al. 2003; He et al. 2004; Wang et al. 2005). Other RCTs have demonstrated favourable effects of acupuncture plus artificial tears compared with artificial tears alone (Nepp et al. 1998; Tseng et al. 2006). However, the number of trials, their total sample size and their methodological quality are too low to draw firm conclusions.

The risk of bias in the studies was assessed based on their descriptions of randomization, blinding, withdrawals and allocation concealment. All the studies were burdened with a high risk of bias. None employed allocation concealment, and one of the six RCTs made an attempt to blind assessors (Tseng et al. 2006). Only two of the RCTs mentioned ethical approvals (Gronlund et al. 2004; Tseng et al. 2006). One (Gronlund et al. 2004) reported details of dropouts and withdrawal and others did not, which may have led to exclusion or attrition biases. Most of the included trials also suffered from a lack of adequate allocation concealment and sufficient sample size, with no power analysis included. Thus, the reliability of the evidence presented here is clearly limited.

Two RCTs (Nepp et al. 1998; Tseng et al. 2006) showed favourable effects of acupuncture plus artificial tears for dry eye compared with artificial tears alone. Because of their design (A + B versus B), these RCTs are unable to demonstrate specific therapeutic effects of the treatment tested. Two of these studies generated positive results, while one RCT failed to do so.

None of the included trials employed a placebo or sham control procedure. Placebo acupuncture is important for identifying the specific effects of acupuncture; however, it is technically impossible to implement blindly (White et al. 2008). A placebo must be indistinguishable from the real treatment, but it must be inert (White et al. 2008). Because there is no such thing as an inactive placebo control for acupuncture, the term ‘sham’ is used instead for any control procedure that is used in a trial to try to make the patient think that they have had treatment (White et al. 2008). For such studies, several sham procedures are available. They range from needling non-acupuncture points, superficially puncturing the skin on acupuncture points and non-penetration on acupuncture points using novel sham needle devices (Streitberger & Kleinhenz 1998; Fink et al. 2001; Park et al. 2002).

The rationale for the selection of acupuncture points was stated in all six RCTs. All of the authors quoted traditional Chinese medicine theory to justify their point selection. Needle stimulation causing a typical needle sensation has been claimed to be important for reaching maximum effects on pain. This needle sensation (de-qi) was considered in one RCT (Tseng et al. 2006), while the other five trials did not report such details. None reported the stimulation and manipulation methods.

Another concern is that ethical approval was reported in only two of the included trials (Gronlund et al. 2004; Tseng et al. 2006). Considering the importance of protecting patients’ rights, acupuncture researchers must develop an awareness of ethical issues similar to their awareness of other aspects of research. Moreover, their reporting of clinical trials should follow CONSORT guidelines (Begg et al. 1996).

Most of the included RCTs originated from China (Pang et al. 2003; He et al. 2004; Wang et al. 2005; Tseng et al. 2006). It is therefore relevant to mention that (Vickers et al. 1998) have shown that not a single acupuncture study from China has ever reported negative results. In our view, this phenomenon further limits the possibility of drawing reliable conclusions from this meta-analysis.

One argument for using acupuncture for the management of dry eye might be that it causes few adverse effects. One RCT (Gronlund et al. 2004) assessed adverse events because of acupuncture treatment and five RCTs (Nepp et al. 1998; Pang et al. 2003; He et al. 2004; Wang et al. 2005; Tseng et al. 2006) did not. No severe adverse effects of acupuncture were noted. Relative to those of standard drug treatments, the adverse effects of acupuncture are mild, infrequent and perhaps even negligible.

Assuming that acupuncture was beneficial for treating dry eye, possible mechanisms of action may be of interest. It has been suggested that acupuncture can influence synthesis and secretion by the lacrimal gland (Gong & Sun 2007). Others have postulated that acupuncture can reduce tension and alleviate pain intensity (or increase pain threshold) (Nepp et al. 2002). However, none of these theories have been confirmed.

The limitations of our systematic review pertain to the potential incompleteness of the evidence reviewed. We aimed at identifying all RCTs on the topic. The distorting effects of publication bias and location bias on systematic reviews are well documented (Ernst & Pittler 1997; Pittler et al. 2000; Rothstein et al. 2005). In the present review, there were no restrictions on the re-view publication language, and a large number of different databases were searched. We are therefore confident that our search strategy located all relevant data on the subject. Further limitations include the paucity and the often suboptimal quality of the primary data.

In conclusion, the results of our systematic review and meta-analyses provide limited evidence for the effectiveness of acupuncture in treating dry eye. However, the total number of RCTs included in the analysis, the total sample size and their methodological quality were too low to draw firm conclusions. Further rigorous RCTs are warranted, but they will need to overcome the many limitations of the current evidence.

Acknowledgements

The authors specially thank Dr J.I. Kim of KIOM for clinical advice for this manuscript. M.S. Lee and T.Y. Choi were supported by the Acupuncture, Moxibustion and Meridian Research Project (K09050) of the Korea Institute of Oriental Medicine.

Ancillary