Description of the condition
Hordeolum is a common inflammation of the eyelid margin. It presents as a red, painful, swollen furuncle with an acute onset and is usually caused by a staphylococcal infection (Mueller 2008; Peralejo 2008; Skorin 2002). The infection can be internal, affecting the meibomian glands, or external, affecting the glands of Zeis or Moll (Wald 2004). External hordeola are known more commonly as styes. In many cases, the lesion drains spontaneously and resolves untreated; however, the infection can spread to other ocular glands or tissues, and recurrences are common. If unresolved, acute internal hordeolum can become chronic or can develop into a chalazion (De Jesus 2004; Hudson 1981; Mueller 2008; Rubin 1995).
Hordeolum is one of the most common diseases of the eye; therefore many people can be affected and many causative factors are known to be related to the disease. Incidence rates for hordeolum are not available because most cases are not reported. Hordeola tend to occur in younger people but are not limited to any age, gender, or racial group (Fuchs 1911; Lederman 1999; Roodyn 1954). Onset is spontaneous and may be related to lid hygiene, an underlying condition, or a systemic infection (Mathew 1966; Wald 2004). Typically, the size of the swelling is a direct indicator of the severity of the infection (Lebensohn 1950). Internal hordeolum tends to be more painful and longer lasting than external hordeolum (Barza 1983; Fuchs 1911; Olson 1991; Wilkie 1956). Cases of recurrent hordeolum are usually the result of failure to eliminate bacteria completely rather than resulting from new infections (Roodyn 1954). Blepharitis (Fuchs 1911; Skorin 2002), acne rosacea (De Jesus 2004), trichiasis, and cicatricial ectropion (Moriarty 1982) are conditions frequently associated with internal hordeolum.
Most cases of internal hordeolum resolve on their own; therefore people with hordeolum often do not seek professional medical treatment (Olson 1991). Home therapies, including heated compresses, lid scrubs, and over-the-counter medications, are often employed without consultation with a medical professional. For times when medical care is sought, a general practitioner or a family physician may be consulted before an ophthalmologist or an optometrist is seen (Fraunfelder 1971; Lebensohn 1950).
Practice standards for the initial treatment of hordeola are conservative, typically limited to the application of warm compresses several times a day, if any treatment is recommended at all (Barza 1983; Fuchs 1911; Olson 1991; Panicharoen 2011; Sethuraman 2009; Wilkie 1956). A topical antibiotic may be prescribed in conjunction with warm compresses (Diegel 1986; Lebensohn 1950; Lederman 1999; Panicharoen 2011; Wald 2004). If the condition is severe and is resistant to topical antibiotics, systemic antibiotics or surgical incision and drainage may be implemented (Moriarty 1982; Mueller 2008; Panicharoen 2011; Rubin 1995; Skorin 2002).
Description of the intervention
Nonsurgical treatments for hordeolum include the application of warm or hot compresses, the use of lid scrubs and digital massage, the administration of antibiotics or steroids, or alternative medicine such as acupuncture and autohemotherapy. Typically, the intent of these interventions is to reduce healing time while relieving the symptoms associated with the lesion. Thus, interventions of interest would be provided during the first week after onset. Beyond one week, it is believed that internal hordeolum may resolve on its own or may require surgical incision and curettage. In addition to resolving the presenting hordeolum, other aims of the interventions are to minimize the risk that the infection may worsen, may spread to other areas, or may become recurrent.
How the intervention might work
The natural history of acute internal hordeolum generally spans one to two weeks, beginning with the appearance of an abscess and concluding with draining of the abscess. Initial treatments for hordeolum have therefore been aimed at promoting the evacuation of pus from the abscess. The application of a warm or hot compress may facilitate drainage by softening the granuloma (Diegel 1986; Fuchs 1911; Moriarty 1982; Skorin 2002). Heated compresses are typically employed for five to 10 minutes several times a day until the hordeolum is resolved.
Lid scrubs consist of mild shampoos or saline solutions and are applied while the affected area is gently massaged. The theory underlying the use of lid scrubs is that they promote lid hygiene and prepare the physical environment for drainage by clearing debris from the lid margin (Driver 1996; Skorin 2002). Creating a clear channel is believed to initiate drainage, similar to the epilation of an eyelash in cases of external hordeolum (Hudson 1981). Also, ingredients used in shampoos break down bacterial membranes, which further decreases the presence of bacteria at the infection site (McCulley 1984). Lid scrubs are commonly recommended in the treatment of other ocular bacterial infections, such as blepharitis, and may prevent the spread of infection (Avisar 1991). In conjunction with lid scrubs, lid massage has been proposed to physically express secretions from the infected glands (Driver 1996; Scobee 1942).
Antibiotics can be administered locally at the site of infection or may be given systemically. Most cases of hordeolum are caused by a staphylococcal species; therefore antibiotics should be effective against the bacteria. Application of topical antibiotics may reduce healing time by fighting against the causative bacterial infection and reducing inflammation. Many topical medications include ingredients that relieve the symptomatic pain of internal hordeolum. Antibiotics can also be applied locally by injection. Systemic antibiotics are sometimes used when local antibiotics are not effective, or when the infection is not localized.
Steroids can be applied topically as ointments or eyedrops. Internal hordeolum has a short course; therefore as little as one steroid treatment could be effective in reducing healing time and relieving symptoms associated with the inflammation (King 1986; Palva 1983).
Why it is important to do this review
Acute internal hordeolum is a common disease experienced by a wide population. Although the course of the disease is relatively short, instances of internal hordeolum are painful and bothersome. Furthermore, improper management of the underlying cause of the infection may lead to recurrent infections or to the development of other disease. Despite the common recommendation to employ heated compresses, their efficacy in treating hordeolum has not been systematically reviewed. If heated compresses are indeed sufficient in treating hordeolum, then more rigorous interventions, such as antibiotics or steroids, may not be warranted for initial treatment. Conversely, comparing the effectiveness and safety of all available interventions, to determine which may be most beneficial to the individual, is also important. A summary of the evidence should assist patients and professionals in determining preferred methods of treatment.
The objective of this review was to investigate the effectiveness and, when possible, the safety of nonsurgical treatments for acute internal hordeolum compared with observation or placebo.
Criteria for considering studies for this review
Types of studies
This review was limited to randomized and quasi-randomized clinical trials. Examples of quasi-randomized allocation include using participants' birth dates, medical record numbers, or order of enrollment to determine treatment groups.
Types of participants
We were interested in studies of participants with a diagnosis of acute internal hordeolum. Studies of participants with only external hordeolum (stye), chronic hordeolum, or chalazia were excluded.
Types of interventions
Nonsurgical interventions were the primary focus of this review. We included trials that compared the use of hot or warm compresses, lid scrubs, antibiotics, or steroids with observation, placebo, or another active intervention for the treatment of acute internal hordeolum. Complementary and alternative therapies, such as acupuncture and bloodletting, were outside the scope of this review.
Types of outcome measures
The primary outcome for this review was the proportion of participants with complete resolution of hordeolum seven days after diagnosis. The seven-day period for resolution was selected because most cases of hordeolum resolve on their own at between one and two weeks. We also analyzed the proportion of participants with complete resolution of hordeolum after 14 days as a secondary outcome, when these data were available.
- The proportion of participants requiring surgical incision and drainage after the treatment period or seven days after diagnosis.
- The incidence of chalazion after the treatment period or seven days after diagnosis.
- The incidence of recurrence of hordeolum after six months and after one year. A recurrent case was considered as any hordeolum that occurred after one month from the resolution of the initial hordeolum and at any location on the same eyelid, or as defined by the included study.
- The incidence of a secondary hordeolum during or after the treatment period or seven days after diagnosis. A secondary hordeolum was defined as a hordeolum that occurred within one month of the initial hordeolum and at a different location than the initial hordeolum, or as defined by the included study.
We planned to report all adverse effects related to the treatment of hordeolum that were reported in the primary studies. Specific adverse outcomes of interest included conjunctivitis; eye irritation; discoloration of the eyelid, conjunctiva, and lens; and corneal damage.
We planned to report economic data.
Quality of life data
We planned to report quality of life data.
Search methods for identification of studies
We searched the Cochrane Central Register of Controlled Trials (CENTRAL) 2012, Issue 7, part of The Cochrane Library. www.thecochranelibrary.com (accessed 26 July 2012), Ovid MEDLINE, Ovid MEDLINE In-Process and Other Non-Indexed Citations, Ovid MEDLINE Daily, Ovid OLDMEDLINE, (January 1950 to July 2012), EMBASE (January 1980 to July 2012), Latin American and Caribbean Literature on Health Sciences (LILACS) (January 1982 to July 2012), the metaRegister of Controlled Trials (mRCT) (www.controlled-trials.com), ClinicalTrials.gov (www.clinicaltrials.gov) and the WHO International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp/search/en). We did not use any date or language restrictions in the electronic searches for trials. We last searched the electronic databases on 26 July 2012. In 2012 we modified the search strategy to include broader terms related to hordeolum.
See: Appendices for details of search strategies for CENTRAL (Appendix 1), MEDLINE (Appendix 2), EMBASE (Appendix 3), LILACS (Appendix 4), mRCT (Appendix 5), ClinicalTrials.gov (Appendix 6) and the WHO International Clinical Trials Registry Platform (ICTRP) (Appendix 7).
Searching other resources
We reviewed the reference lists from potentially eligible studies to identify further studies. In addition we used the Science Citation Index to search for references that cited potentially eligible studies (last searched on 11 Oct 2012 with no relevant studies identified).
Data collection and analysis
Selection of studies
Two review authors independently assessed the titles and abstracts from the electronic literature searches and the manual search to identify possible trials of interest according to the 'Criteria for considering studies for this review'. We designated each reference identified from the searches as (a) relevant, (b) possibly relevant, or (c) not relevant for this review. We retrieved full text copies of the articles if an abstract was classified as (a) or (b). Each article was then independently assessed by two people and was classified as (1) include in review, (2) awaiting classification, or (3) exclude from review. We resolved discrepancies between authors by consensus.
Data extraction and management
As no studies were identified for inclusion in this review, no data extraction or assessment of risk of bias was performed. If, in the future, relevant studies become available, we will undertake the following methods for updating this review.
Two review authors will independently extract data using the data extraction forms created by the Cochrane Eyes and Vision Group. For each included study, we will extract data on study characteristics, interventions, outcomes, cost, quality of life, and other relevant information. One review author will enter the data into Review Manager (RevMan 2012) and a second review author will verify the data entry. Discrepancies between review authors will be resolved by the third review author.
Assessment of risk of bias in included studies
Two review authors will independently assess the risk of bias of included studies based on the methods provided in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Sources of potential bias affecting the quality of a study will be divided into six domains that include:
- Selection bias: sequence generation and allocation concealment.
- Performance bias: masking (blinding) of participants and study personnel.
- Detection bias: masking (blinding) of outcome assessors.
- Attrition bias: incomplete outcome data and rates of missing data among treatment groups.
- Reporting bias: selective outcome reporting.
- Other sources of bias: funding source and other potential sources of bias.
For every study included in the review, we will assess each domain grouping to have (a) low risk of bias, (b) unclear or not reported risk of bias, or (c) high risk of bias. Discrepancies between review authors will be resolved by a third review author. For studies classified as unclear or not reported, we will contact the authors of the study for further information in an attempt to reclassify the quality of the study. If no informative response is received within 6 weeks, we will assess the study on the basis of available information.
Measures of treatment effect
The measures of treatment effect will depend on the types of data presented in the included studies and will be identified by the definitions given in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2011).
The primary outcome of interest, the proportion of participants with complete resolution of hordeolum at seven days after diagnosis, will be analyzed as a dichotomous variable: resolved versus not resolved. Data on the proportion of participants requiring surgical incision and drainage after treatment, the proportion of participants developing a chalazion after treatment, the proportion of participants with recurrent hordeola, and the number of secondary hordeola will also be analyzed as dichotomous data. We will report dichotomous data as a summarized risk ratio with 95% confidence interval.
We will report continuous data as a weighted mean difference with its standard deviation. We anticipate that available economic and quality of life data will be analyzed as continuous data.
We will summarize ordinal data qualitatively.
Counts and rate data
We will summarize counts and rate data in rate ratios when the event is rare, and as continuous outcome data when the event is more common. We will analyze adverse events data as counts and rates.
Unit of analysis issues
The unit of analysis for this review will be an eyelid of an individual participant.
Dealing with missing data
We will contact authors of included studies in an attempt to obtain missing data. We will set the response time at six weeks and will document any communications with study authors. If data cannot be retrieved, we will impute data from the reported data in the study. We will report loss to follow-up when available.
Assessment of heterogeneity
We will test for statistical heterogeneity using the I
Assessment of reporting biases
We will use funnel plots to assess the possibility of reporting biases if more than ten studies are available.
If limited heterogeneity is suggested (defined here as I
Subgroup analysis and investigation of heterogeneity
We will investigate heterogeneity by conducting subgroup analyses provided sufficient information is available. Subgroups of interest include sex, age, use of contact lenses, including soft lenses versus hard lenses, and the frequency of hordeolum occurrences, coinfections, and other comorbidities at baseline.
We will investigate the impact of studies with a high likelihood of bias, or with missing data, as well as the impact of unpublished studies, using sensitivity analyses.
Description of studies
Results of the search
A total of 517 references were identified by the electronic searches as of 21 June 2010. After the titles and abstracts were screened, 19 references were classified as being potentially relevant. All of these 19 references, which reported 18 unique studies, were excluded.
For the update of this review, we identified 427 additional references from electronic searches as of 26 July 2012. After the titles and abstracts were screened, six references were classified as being potentially relevant. All of these six references, which reported five unique studies, were excluded.
Overall 23 studies were excluded in this review. The reasons for exclusion are described in the table 'Characteristics of excluded studies'.
Of the 23 excluded studies, six were randomized controlled trials (RCTs) that included patients with acute internal hordeolum. The first included pediatric participants with lid inflammation and was conducted to evaluate the safety of loteprednol etabonate 0.5% and tobramycin 0.3% ophthalmic suspension (Zylet
Risk of bias in included studies
No studies were included in this review, thus no risk of bias assessment was done.
Effects of interventions
No studies were included in this review, thus no effects of interventions were reported.
Summary of main results
No trials were identified for inclusion in this review.
Overall completeness and applicability of evidence
Most of the references identified from the search for this review were related to external hordeola (styes) or chalazion. By and large, the few references specific to acute internal hordeolum either reported recommendations for treatment without cited evidence or were reports of interventional case series, case studies, or other types of observational study designs. The only clinical trials found that included participants with acute internal hordeolum were not eligible for the review because they included multiple conditions and did not stratify by specific diagnoses included participants who underwent surgical treatment as a criterion for study enrollment, or evaluated treatments that were not under the scope of review. Furthermore, the bulk of the literature was published more than 20 years ago.
Potential biases in the review process
The primary source of bias for this review was anticipated to be selection bias, specifically, the identification and inclusion of relevant studies. Before beginning the review process, we expected that few trials had been published on hordeolum, that various authors used different terminologies when referring to different classifications of hordeola (i.e. hordeolum, stye, chalazia, etc.), and that relevant studies may be obtained from older publications. We therefore designed a broad search strategy for the electronic databases to facilitate identification of potentially relevant studies. We also manually searched the reference lists of potentially relevant studies to identify older studies that may not be included in electronic databases.
To minimize bias during the process of selecting studies for this review, two review authors screened the references from the electronic search and independently classified them for inclusion or exclusion. We included potentially relevant references that mentioned any type of hordeolum or external eye inflammation for assessment at the full text level. Inclusion and exclusion were determined by using the definition of the disease given in the full text article. Furthermore, one review author screening the studies had a clinical background (JN), and one had a methodological background (KL).
Agreements and disagreements with other studies or reviews
Although it is the most recommended therapy for hordeola, the application of warm compresses has not been shown to be effective in accelerating healing time or reducing symptoms associated with hordeolum in a controlled trial. Moreover, no evidence indicates that warm compresses alone would eliminate the infection. It is also unclear whether medical treatment or lid hygiene is effective in treating acute internal hordeolum.
Implications for practice
Common interventions for the treatment of acute internal hordeolum include warm compresses applied at home, topical medications and lid scrubs available over-the-counter, antibiotics or steroids, lid massages, and others. At this time evidence regarding the effectiveness of these nonsurgical interventions for treating acute internal hordeolum is insufficient. Clinical practice decisions should be based on physician judgment, and available treatment options should be discussed with patients.
Implications for research
Generally, RCTs are considered the gold standard for comparing the efficacy of interventions. However, because of the relative mildness and short duration of the disease, study participants may be limited to more severe cases that are not representative of the general population; recruitment of participants at onset may be challenging. Even with these considerations, controlled clinical trials would be useful in determining which interventions are effective for the treatment of acute internal hordeolum.
We thank Iris Gordon and Lori Rosman for devising and implementing the electronic search strategy for the review. We thank the Cochrane Eyes and Vision Group editorial team for assisting with preparation of the protocol and review. We also thank Barbara Hawkins, Karen Blackhall, Daniel Ezra, and John Bladen for their comments; Takeshi Iwase and Sueko Matsumura for their assistance with evaluating Japanese-language articles; and Tsung Yu and Xue Wang for their assistance with evaluating Chinese-language articles. We especially thank Nancy Fitton for her work on writing the Plain Language Summary.
Data and analyses
This review has no analyses.
Appendix 1. CENTRAL search strategy
#1 MeSH descriptor Hordeolum explode all trees
#2 (Hordeol* or stye or styes)
#3 MeSH descriptor Eye explode all trees
#5 (#3 AND #4)
#6 MeSH descriptor Meibomian Glands explode all trees
#7 (Meibomian* adj3 (gland* or cyst* or infection* or inflammat*))
#8 (tarsal adj3 (gland* or cyst* or infection* or inflammat*))
#9 (palpebral adj3 (gland* or cyst* or infection* or inflammat*))
#10 (conjunctiv* adj3 (gland* or cyst*))
#11 (gland* adj5 (zeis* or Moll*))
#12 (lid* or eyelid* or "eye margin") adj3 inflammat*
#13 (lid* or eyelid* or "eye margin") adj3 infection*
#14 (#1 OR #2 OR #4 OR #6 OR #7 OR #8 OR #9 OR #10 OR #11 OR #12 OR #13)
Appendix 2. MEDLINE (OvidSP) search strategy
1. Randomized Controlled Trial.pt.
2. Controlled Clinical Trial.pt.
3. (randomized or randomised).ab,ti.
5. drug therapy.fs.
9. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8
10. exp animals/ not humans.sh.
11. 9 not 10
12. exp hordeolum/
13. (Hordeol* or stye or styes).tw.
15. exp eyes/
16. 14 and 15
17. exp meibomian glands/
18. (Meibomian* adj3 (gland* or cyst* or inflammat* or infection*)).tw.
19. (tarsal adj3 (gland* or cyst* or inflammat* or infection*)).tw.
20. (palpebral adj3 (gland* or cyst* or inflammat* or infection*)).tw.
21. (conjunctiv* adj3 (gland* or cyst*)).tw.
22. (gland* adj5 (zeis* or Moll*)).tw.
23. ((lid* or eyelid* or eye margin) adj3 inflammat*).tw.
24. ((lid* or eyelid* or eye margin) adj3 infection*).tw.
25. 12 or 13 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24
26. 11 and 25
The search filter for trials at the beginning of the MEDLINE strategy is from the published paper by Glanville et al (Glanville 2006).
Appendix 3. EMBASE.com search strategy
#1 'randomized controlled trial'/exp
#3 'double blind procedure'/exp
#4 'single blind procedure'/exp
#6 #1 OR #2 OR #3 OR #4 OR #5
#7 'animal'/exp OR 'animal experiment'/exp
#9 #7 AND #8
#10 #7 NOT #9
#11 #6 NOT #10
#12 'clinical trial'/exp
#13 (clin* NEAR/3 trial*):ab,ti
#14 ((singl* OR doubl* OR trebl* OR tripl*) NEAR/3 (blind* OR mask*)):ab,ti
#18 'experimental design'/exp
#19 'crossover procedure'/exp
#20 'control group'/exp
#21 'latin square design'/exp
#22 #12 OR #13 OR #14 OR #15 OR #16 OR #17 OR #18 OR #19 OR #20 OR #21
#23 #22 NOT #10
#24 #23 NOT #11
#25 'comparative study'/exp
#27 'prospective study'/exp
#28 control*:ab,ti OR prospectiv*:ab,ti OR volunteer*:ab,ti
#29 #25 OR #26 OR #27 OR #28
#30 #29 NOT #10
#31 #30 NOT (#11 OR #23)
#32 #11 OR #24 OR #31
#34 hordeol*:ab,ti OR stye:ab,ti OR styes:ab,ti
#37 #35 AND #36
#38 'meibomian gland'/exp
#39 (meibomian* NEAR/3 (gland* OR cyst* OR inflammat* OR infection*)):ab,ti
#40 (tarsal NEAR/3 (gland* OR cyst* OR inflammat* OR infection*)):ab,ti
#41 (palpebral NEAR/3 (gland* OR cyst* OR inflammat* OR infection*)):ab,ti
#42 (conjunctiv* NEAR/3 (gland* OR cyst*)):ab,ti
#43 (gland* NEAR/5 (zeis* OR moll*)):ab,ti
#44 ((lid* OR eyelid* OR 'eye margin') NEAR/3 inflammat*):ab,ti
#45 ((lid* OR eyelid* OR 'eye margin') NEAR/3 infection*):ab,ti
#46 #33 OR #34 OR #37 OR #38 OR #39 OR #40 OR #41 OR #42 OR #43 OR #44 OR #45
#47 #32 AND #46
Appendix 4. LILACS search strategy
Hordeol$ OR stye OR styes OR sty OR Meibomian$ OR (Tarsal gland$ OR tarsal cyst$ OR tarsal inflammat$ or tarsal infection$ OR palpebral gland$ OR palpebral cyst$ OR palpebral inflammat$ OR palpebral infection$) OR ((conjunctiv$) AND (gland$ OR cyst$)) OR ((zeis$ OR moll$) AND (gland$)) OR (Eyelid$ AND (inflammat$ OR infection$))
Appendix 5. metaRegister of Controlled Trials search strategy
Hordeolum OR hordeola OR stye OR styes OR sty OR Meibomian OR Eyelid inflammation
Appendix 6. ClinicalTrials. gov search strategy
Hordeolum OR hordeola OR stye OR styes OR sty OR Meibomian OR Eyelid inflammation OR Tarsal gland OR tarsal cyst OR tarsal inflammation or tarsal infection OR palpebral gland OR palpebral cyst OR palpebral inflammation OR palpebral infection OR conjunctiva gland OR conjunctiva cyst
Appendix 7. WHO International Clinical Trials Registry Platform search strategy
Hordeolum OR hordeola OR stye OR styes OR Meibomian OR Eyelid inflammation OR Tarsal gland$ OR tarsal cyst$ OR tarsal inflammation or tarsal infection$ OR palpebral gland$ OR palpebral cyst$ OR palpebral inflammation OR palpebral infection$ OR conjunctiva gland$ OR conjunctiva cyst$
Last assessed as up-to-date: 26 July 2012.
Contributions of authors
Conceiving the review: KD, KL
Designing the review: KL, JN
Co-ordinating the review: KL
Data collection for the review:
- Designing electronic search strategies: Cochrane Eyes and Vision Group
- Undertaking electronic searches: Cochrane Eyes and Vision Group
- Screening search results: KL, JN
- Organizing retrieval of papers: KL
- Screening retrieved papers against inclusion criteria: KL, JN
- Appraising risk of bias of papers: KL, JN
- Extracting data from papers: KL, JN
- Writing to authors of papers for additional information: KL
- Providing additional data about papers: KL, JN
- Obtaining and screening data on unpublished studies: KL, JN
Data management for the review:
- Entering data into RevMan: KL, JN
Analysis of data: KL, JN, KD
Interpretation of data:
- Providing a methodological perspective: KL, KD
- Providing a clinical perspective: JN
- Providing a policy perspective: JN
Writing the review: KL, JN, KD
Providing general advice on the review: KL, JN, KD
Securing funding for the review: KD
Updating the review:
- KL reviewed search results, extracted data, and wrote the review
- JN and KD revised and edited the review
We acknowledge Sueko Matsumura for assistance in screening search results for the update.
Declarations of interest
Sources of support
- No sources of support supplied
- Grant 1 U01 EY020522-01, National Eye Institute, National Institutes of Health, USA.
Medical Subject Headings (MeSH)
MeSH check words