Assessing the value of the water drinking test in glaucoma—a systematic review and meta-analysis

Abstract

This systematic review and meta-analysis examines the relationship between intraocular pressure (IOP) parameters derived from the water drinking test (WDT) and diurnal IOP monitoring, and evaluates the reproducibility of the WDT and its association with future glaucomatous visual field (VF) loss. A literature search was performed on PubMed, EMBASE, and Cochrane Library from inception to 31 March 2023. Cohort, cross-sectional and observational studies reporting WDT results in glaucoma patients were included. Meta analysis with random-effect model was performed using “metafor” package in R version 3.2.1. 641 studies were identified on initial literature search. 38 studies (2479 subjects) were included in final analysis. Meta-analytic estimates of 5 studies (310 subjects) found strong positive correlation in peak IOP between the WDT and diurnal IOP monitoring (r = 0.92, 95% CI = 0.75, 1.08, p < 0.0001). However, there was only weak positive correlation for IOP fluctuation between both methods (r = 0.26, 95% CI = 0.06,0.47, p < 0.0001). Meta-analytic estimates of 3 studies (189 subjects) suggested a trend of the diurnal peak IOP being lower than that derived from the WDT (MD = −2.37 mmHg, 95% Limit of Agreement (LOA) =−5.58,0.84, p = 0.147). Two studies found that a higher WDT peak IOP was associated with greater future VF progression. Two studies demonstrated good reproducibility in peak IOP measurements for WDTs conducted across different days. In conclusion, there was a strong positive correlation between IOP peak measurements from the WDT and diurnal IOP monitoring in glaucoma patients. The WDT peak IOP demonstrated good reproducibility and may be associated with greater future VF progression.

Introduction

Glaucoma is the leading cause of irreversible blindness worldwide [1], and is expected to affect 111.8 million people by 2040 [2]. While its underlying pathogenesis is likely multifactorial, intraocular pressure (IOP) remains the primary modifiable risk factor for glaucoma, and IOP reduction is the mainstay of glaucoma treatment today [3]. As such, accurate evaluation and monitoring of IOP are crucial in guiding treatment decisions to prevent disease progression.

Diurnal IOP curves provide valuable information about IOP changes throughout the day, which have been shown to influence glaucoma progression [4] and are at times used to aid the diagnosis and monitoring of glaucoma [5]. However, 24-hour diurnal IOP is usually monitored by phasing in the clinic or by home monitoring devices, which are resource-intensive or demand compliance and cooperation from patients.

The water drinking test (WDT) has been proposed as a potential alternative tool to assess IOP dynamics in glaucoma patients. The WDT involves ingesting a standardized volume of water, typically 800–1000 mL, followed by interval IOP measurements over a defined period, typically an hour. The basis of the WDT rests on the understanding that ingestion of a significant volume of fluid within a short time tests the eye’s outflow facility and assesses the eye’s ability to cope with the associated increased IOP [6]. Initially explored as a diagnostic test in the 1900s [7], the WDT was found to have generally low specificity and sensitivity in diagnosing glaucoma [6]. However, in recent years it has been utilized as a provocative test to estimate one’s outflow facility reserve, through derived measures such as IOP instability and peak IOP. Peak IOP values, determined by the WDT, has also been used to predict future glaucoma progression and to compare outcomes of pharmacological and surgical interventions for glaucoma [6].

Despite the extensive literature available, there have been few attempts to systematically consolidate and quantitatively evaluate the application and value of the WDT in glaucomatous eyes, with most prior reviews being narrative in nature [6, 8]. Our study is, to the best of our knowledge, the first systematic review and meta-analysis on this topic and aims to quantitatively examine the relationship between the WDT results and diurnal IOP curve results, as well as to qualitatively evaluate the clinical utility and reliability of the WDT in various aspects of glaucoma management.

Methodology

This systematic review and meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist and guidelines [9] (Appendix 1). Institutional review board approval was not required as this study utilized published data available in the public domain and did not use individual-level data. All research adhered to the tenets of the Declaration of Helsinki.

Search strategy

Literature search was performed in the MEDLINE, EMBASE, and Cochrane Library databases from inception up to 01 March 2023. Key search terms included: “water drinking”, “glaucoma”, “intraocular pressure”, “diurnal IOP”, “visual field” and their synonyms. The detailed search strategy is available in Appendix 2. Reference lists of manuscripts were hand searched for further relevant articles. Two reviewers (EJ, CXYG) independently performed the literature search, with title and abstract screening before cross-checking their list of identified articles. Differing opinions were adjudicated by the senior author (BCHA).

Eligibility criteria

This review included retrospective and prospective cohort studies, case-control studies, and cross-sectional studies written primarily in English or Chinese that reported on the efficacy of the WDT. Only studies including participants with a definitive diagnosis of glaucoma (diagnosed by a trained professional based on evidence of structural and/or functional impairment), without restriction on the type of glaucoma were included in the review. We excluded the following: (1) non-English studies where the corresponding English translation of the full-text was not available; (2) editorials, correspondences, non-human studies, letters, reviews, conference abstracts, case series and case reports; (3) studies published before the year 2000; (4) studies with a sample size of <20 participants (5) participants aged <18 years old; (6) participants with concomitant non-glaucomatous optic neuropathy, retinal disease or intraocular inflammation that might affect visual field (VF) or IOP measurements, or with a history of ocular trauma; (7) studies including participants with suspected glaucoma or ocular hypertension without a definitive glaucoma diagnosis.

Data collection and risk of bias assessment

The following information was manually extracted from the full texts of included studies – (1) study details, demographics, type of glaucoma and baseline treatments; (2) WDT protocols and diurnal curve measurement methods; (3) WDT results and diurnal IOP curves (including quantitative measures of correlation and agreement for three IOP parameters – mean IOP, peak IOP and IOP fluctuation); (4) quantitative data on the relationship between WDT results and VF parameters indicating glaucoma severity and progression; (5) qualitative and quantitative data on factors affecting WDT results; (6) WDT results before and after various glaucoma treatment modalities. Two reviewers independently extracted information (EJ, CXYG) in a pre-defined template and compared results to ensure accuracy in data collection. The risk of bias of included studies was independently assessed by two reviewers (EJ, CXYG) using the Newcastle-Ottawa Scale (NOS) [10] for non-randomized studies and a modified NOS for cross-sectional studies. Studies were assessed based on 3 categories, namely selection, comparability, and outcome. Differing opinions were adjudicated by the senior author (BCHA).

Data synthesis and analysis

Our primary outcomes included the pooled correlation coefficients between the WDT and the diurnal IOP curve, for two IOP parameters – (1) peak IOP and (2) IOP fluctuation. Correlation coefficients were pooled using a random-effects model, given the expected heterogeneity across studies. Where necessary, Spearman’s correlation coefficients were converted to an approximate Pearson’s coefficient [11], and Fisher transformation [12] was used before the Pearson’s coefficients were pooled.

While the interpretation of what constitutes a ‘high’ or ‘low’ correlation coefficient may vary depending on the context and field of study [13], we used the following: (1) Between ±0.90 and ±1.00 = Very High; (2) Between ±0.70 and ±0.90 = High; (3) Between ±0.50 and ±0.70 = Moderate; (4) Between ±0.30 and ±0.50 = Low; (5) Between ±0.00 and ±0.30 = Negligible.

Because of insufficient data in current literature, we were unable to conduct a pooled meta-analysis for our secondary outcomes. However, we discussed these qualitatively. Our secondary outcomes included—(1) the relationship between IOP parameters of the WDT and long-term IOP curve; (2) the ability of WDT results to predict glaucoma severity and future glaucoma progression; (3) factors affecting WDT results; (4) WDT results before and after various glaucoma treatments; (5) the reproducibility of WDT results.

We also gathered data on the agreement between peak IOP, mean IOP and IOP fluctuation of the WDT and diurnal IOP curve via the mean difference and 95% Limits of Agreement (LOA) of the Bland Altman plots. The Standard Deviation (SD) was then calculated using (upper LOA-mean)/1.96. The mean difference and SD of the Bland Altman plots were then pooled using the random effects model, as previously done by Joosten et al. [14].

Statistical analysis was done using Meta-Mar (1.1.0) [15] and the metafor package from R 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria). Publication bias was assessed through a comparison-adjusted funnel plot and the Egger’s linear regression test.

For assessment of the presence of heterogeneity within each pairwise comparison, the Cohran’s Q and I² statistics were used. I² threshold values of 25%, 50% and 75% signified low, moderate, and large amounts of heterogeneity respectively. A 2- sided p-value of less than 0.05 was regarded as statistically significant. To test the robustness of our results and explore sources of heterogeneity, we also performed a sensitivity analysis, by excluding studies whose CI did not include the mean of our pooled results and then re-running the analysis afterwards.

Results

Summary of included studies

A total of 641 publications were identified from initial literature search, of which 38 studies [16–53] were included in final analysis (Fig. 1). These comprised 5 studies from Asia, 1 study from Africa, 3 studies from Europe, 1 study from North America, 20 studies from South America, 6 studies from Middle East and 2 studies from Oceania. Most studies (30/38) were prospective in nature and only 8 were retrospective [16, 19, 23, 43, 45, 48, 49, 51]. The 38 studies included a total of 2479 participants. In this pooled cohort, the mean age of participants was 59.8 years, similar to the general demographic of glaucoma patients [54]. The most common form of glaucoma reported was open-angle glaucoma (OAG) with 1843 participants (74.3%), followed by exfoliative glaucoma (XFG) with 194 participants (7.83%).

Fig. 1.

PRISMA flowchart.

With regards to the WDT protocol, almost half of all studies (17/38) included a 2–3 h fast before the test, followed by consumption of water within 5–20 min. 11 studies had their participants consume 1 L, while in 8 studies, participants consumed 800 ml of water. The IOP was measured every 15 min for 1 h by an optometrist or ophthalmologist using a Goldmann Applanation Tonometer (GAT). For diurnal IOP curves, most of the studies were conducted in clinic during office hours (8am–6 pm) using the GAT, while in 2 studies [32, 44], IOPs were measured over 24 h. Details of the included studies are summarized in Table 1. Results of the risk of bias according to the Newcastle Ottawa Scale (NOS) [17] are displayed in Appendix 3.

Table 1.

Characteristics of Included Studies.

S/N	Author & Year	Country (city)	Study design	Number of Participants (Male / Female)	Age, Mean (SD)	Type of Glaucoma (n=number of patients)	Medication Washout Performed (Y/N/NA)	Tonometry	Fasting Period	WDT Protocol	Diurnal IOP Curve Measurement Protocol	Definition of IOP Fluctuation
1	Carolina Nicolela Susanna, 2022	Brazil	Retrospective cross-sectional study	49 (26/23)	60 (12)	POAG(49)	N	GAT	2 h	800 mL in 5 min	NA	Not Applicable
2	Sujatha V Kadambi, 2021	India (Chennai)	Prospective observational study	100 (22/78)	55.66 (14.58)	EYES: PACG(28), NTG(31), POAG(56)	Not stated if on medications	GAT	2 h	10 mL/kg in 5-15 min	IOP taken at 2–3 h intervals from 8AM - 3PM.	Not Available
3	Pelin Özyol, 2016	Turkey	Prospective observational study	32 (19/13)	POAG:54.5 (7.28) Healthy:51.2 (9.81)	POAG(20)	At baseline: No anti-glaucoma medications At 6 weeks post prostaglandin treatment: N	GAT	2 h	1000 mL in 5 min	NIL	Peak IOP-baseline IOP
4	M. Reza Razeghinejad, 2018	Iran	Interventional case series	203 (90/113)	54 (18)	POAG: 103 (50.7) PACG: 40 (19.7) XFS: 24(11.8) Congenital: 8 (3.9) Others: 28 (13.8)	N	NCT	3 h	1000 mL in 5 min For those under 16: 15 mL/kg in 5 min	NA	Peak IOP-baseline IOP
5	Izabela Almeida, 2021	Brazil	Prospective observational study	63 (37/26)	60.7 (11.8)	Stable OAG (63)	N	GAT	2 h	800 mL in 5 min	NA	Peak IOP-baseline IOP
6	Erhan Ozyol, 2016	Turkey (Mugla)	Prospective observational study	64 (38/26)	65.6 (10.7)	XFS(34), XFG(30)	N	GAT	2 h	1000 mL in 5 min	NA	Peak IOP-baseline IOP
7	Carlos Gustavo Vasconcelos-Moraes, 2008	Brazil	Prospective intervention study	97 (59/38)	43.2 (6.4)	POAG (97)	Y	GAT	Not Available	1000 mL in 5 min	IOP measurements at 8 am, 10 am, 12 pm, 2 pm, and 4 pm.	Not Applicable
8	Remo Susanna Jr, 2006	Brazil (Sao Paulo)	Retrospective cross-sectional study	101 (NA)	NA	POAG(101)	N	GAT	3 h	1000 mL in 5 min	NA	Peak IOP-baseline IOP
9	Michele Vetrugno, 2005	Italy	Prospective intervention study	280 (114/166)	Timolol: 65.2(7.1) Dorzolamide: 65.6 (7.4) Brinzolamide: 67.3 (11.2) Latanoprost: 66.4 (12.5) Travoprost: 61.8 (6.8) Bimatoprost: 66 (8.9) Brimonidine: 63.1 (5.7)	POAG(280)	N	GAT	Empty stomach	1000 mL in 10 min	NA	Not Applicable
10	Patricia Martinez, 2017	United States (Philadelphia)	Prospective intervention study	34 (18/22)	Trabeculectomy:69.7 (12.75), Tube Shunt Group:66.2 (11.68)	OAG(34)	N	GAT	Unstated	10 mL/kg in 15 min	NA	Not Applicable
11	Flavio Mac Cord Medina, 2009	Brazil	Prospective observational study	45 (24/21)	POAG:55.7 (12.3), Healthy:53.7 (13.4)	POAG(15)	Y	GAT	6 h	1000 mL in 5 min	NA	Peak IOP-baseline IOP
12	Chen Chiao Hong, 2000	Taiwan	Prospective observational study	56 (16/40)	Trabeculectomy: 48.2 (12.6), Healthy:44.6 (10.2)	PACG(11), POAG(26), Neovascular (2), Traumatic (3)	N	GAT	Unstated	1500 mL in 5 min	NA	Not Applicable
13	Danesh-Meyer Helen V., 2008	New Zealand	Prospective observational study	60 (28/32)	Surgical:70 (9), Medical:68 (12)	OAG(60)	N	GAT	2 h	1000 mL in 5 min	NA	Not Applicable
14	Renato Antunes Schiave Germano, 2021	Brazil	Prospective interventional longitudinal, and randomized clinical trial	29 (18/11)	56.6 (11.5)	POAG(29)	Y	GAT	2 h	800 mL in 5 min	NA	Not Applicable
15	Mansouri, K, 2008	Switzerland (Lausanne + Basel)	Multicenter prospective interventional study	60 (32/28)	Trabeculectomy: 67.1 (9.24), DSCI: 72.5 (11.9), Latanoprost: 71.2 (10.4)	POAG(60)	N (Patients were advised not to take latanoprost during the study hours)	GAT	Unstated	1000 mL in 15 min	4 IOP measurements taken at 3 h intervals from 08:00 to 17:00.	Not Applicable
16	Lourenco Adriana Sobral, 2021	Brazil (Brasília)	Prospective observational study	40 (24/16)	Pseudophakic: 76.00 (12.31) Phakic: 67.50 (6.24)	POAG(40)	N	GAT	2 h	1000 mL in 5 min	5 IOP taken at 2 h intervals from 08:00 to 16:00	Not Applicable
17	Feng, H, 2023	China (Shen Zhen)	Prospective observational study	87 (39/49)	Median:34.5, Range: 19-69	HTG(33), NTG(28)	NA	NCT	Unstated	1000 mL in 5 min	12 IOP taken at 2 h intervals for 24 h from 08:00	Not Available
18	Phu Jack, 2021	Australia (New South Wales)	Prospective Cross-sectional study	27 (11/16)	Median:57, IQR:(48.5–66.5, 36–83)	POAG(7)	NA (no anti-glaucoma medications)	GAT / PAT	2 h	10 mL/kg in 5 min	7 days of 4 IOP measurements a day from 05:00–01:00 the next day	Not Applicable
19	De Moraes Carlos G., 2017	Brazil	Prospective longitudinal study	96 (45/51)	67.8 (10.9)	POAG(144 eyes)	N	GAT	Unstated	800 mL in 5 min	NA	Peak IOP-baseline IOP
20	Hatanaka Marcelo, 2016	Brazil	Prospective cross-sectional observational study	45 (31/14)	POAG:62.19 (12.23), Normal:57.5 (13.17)	POAG(31)	N	GAT	Unstated	1000 mL in 5 min	NA	Peak IOP-baseline IOP
21	Caiado Rafael Ramos, 2014	Brazil (Sao Paulo)	Prospective observational study	60 (34/26)	POAG:61.5 (5.6), Healthy:59.7 (8.0)	POAG(30)	N	GAT	Fasted, duration unspecified	1000 mL in 5 min	SDTC in the sitting and supine positions. 5 IOP measurements taken at 2 h intervals from 08:00 to 16:00	Highest IOP-Lowest IOP
22	Firat Penpe Gu ¨l, 2021	Turkey	Prospective observational study	42 (18/24)	66.95 (6.80)	XFG(42)	N	GAT	3 h	1000 mL in 5 min	5 IOP measurements taken at 2 h intervals from 08:00 to 16:00	Fluctuation = Peak IOP-baseline IOP % IOP fluctuation= (IOP peak – baseline IOP)/baseline IOP x100 for WDT, (IOPpeak – baseline IOP/lowest IOP) x 100 for mDTC
23	Olatunji, Olayemi P., 2020	Nigeria (Eleta)	Prospective interventional comparative study	50 (27/23)	59.5 (8.2)	POAG(50)	NA (untreated because newly diagnosed)	GAT	Unstated	800 mL in 5 min	5 IOP measurements taken at 2 h intervals from 07:00 to 15:00	Not Available
24	Scoralick Ana Luiza Bassoli, 2019	Brazil	Prospective observational study	87 (52/35)	61.9 (12.5)	POAG(87)	N	GAT	2 h	800 mL in 5 min	NA	Not Available
25	Poon, Yi-Chieh, 2016	Taiwan	Prospective observational study	30 (18/12)	62.0 (7.8)	PACG(15), PAOG(15)	N	Tonopen (TONO-PEN XL, Reichert Inc., Depew, NY USA)	2 h	10 mL/kg in 10 min	NA	Peak IOP-baseline IOP
26	Mocan, Mehmet C, 2016	Turkey	Prospective cross-sectional study	75 (30/45)	XFG:67.7 (7.5), XFS:68.6 (7.1), Control:65.7 (6.7)	XFG(25)	N	GAT	2 h	1000 mL in 5 min	2 IOP measurements taken at 09:00 and 17:00. The diurnal IOP fluctuation was defined as the difference between these 2 values	Fluctuation = Peak IOP-baseline IOP % IOP fluctuation= (IOP peak – baseline IOP)/baseline IOP x100
27	Kanadani, Fabio N., 2016	Brazil (Minas Gerais)	Prospective comparative study	79 (NA)	NA	“Early glaucoma”(MD < –6 dB, glaucomatous optic neuropathy):(31)	NA (no anti-glaucoma medications)	Not stated for WDT	Fasted, duration unspecified	1000 mL in 5 min	NA	Not Applicable
28	Babic, Mirko, 2015	Brazil (Sao Paulo)	Retrospective cohort study	34 (NA)	NA	POAG(34)	N	GAT	2 h	800 mL in 5 min	NA	Peak IOP-baseline IOP
29	Rei Sakata, 2013	Japan (Tokyo, Gifu)	Prospective observational study	33 (18/15)	50.3 (11.3)	NTG (66)	NA (no anti-glaucoma medications)	GAT	4 h	14 mL/kg in 5 min	8 IOP measurements taken at 3 h intervals from 09:00 to 06:00 the next day.	Peak IOP-baseline IOP
30	Rafael Lacerda Furlanetto, 2010	Brazil (Sao Paulo)	Retrospective case series	55 (35/20)	65.65 (28.28)	POAG (55)	NA (no anti-glaucoma medications)	GAT	Unstated	1000 mL in 5 min	NA	%IOP fluctuation = (IOP peak – baseline IOP)/baseline IOP x100
31	Carlos Gustavo V De Moraes, 2009	Brazil (Sao Paulo)	Prospective cohort study	22 (13/9)	54.3 (8.2)	POAG (22)	N	GAT	4 h	1000 mL in 5 min	NA	%IOP fluctuation = (IOP peak – baseline IOP)/baseline IOP x100
32	Verônica C. Lima, 2008	Brazil (Sao Paulo)	Prospective observational study	39 (10/29)	68.5 (11.9)	POAG (39)	N	Not stated	4 h	1000 mL in 5 min	NA	Peak IOP-baseline IOP
33	R Susanna, Jr, 2005	Brazil (Sao Paulo)	Retrospective observational study	76 (35/41)	67.1 (11.8)	POAG (76)	N	GAT	Unstated	1000 mL in 5 min	NA	%IOP fluctuation = (IOP peak – baseline IOP)/baseline IOP x100
34	Felipe A Medeiros, 2002	Brazil (Sao Paulo)	Retrospective cohort study	60(NA)	Medical group: 67 (12) Surgical group: 63 (11)	POAG (60)	N	GAT	4 h	1000 mL in 5 min	4 IOP measurements at 3 h intervals from 08:00 - 17:00	Peak IOP-baseline IOP
35	Hatanaka Marcelo, 2010	Brazil (Sao Paulo)	Prospective, open-label, randomized controlled clinical trial	49(NA)	63.22 (11.49)	POAG (98)	N (wash out only for screening visit)	GAT	Unstated	1000 mL in 5 min	4 IOP measurements during the day (8:00 AM, 10:30 AM, 2:00 PM, and 4:00 PM)	Not Applicable
36	Renato A.S. Germano, 2016	Brazil (Sao Paulo)	Retrospective interventional cohort study.	41 (25/16)	69 (12)	OAG (41)	Y (for the first WDT only)	GAT	2 h	800 mL in 5 min	NA	Not Applicable
37	Bianca N. Susanna, 2022	Brazil (Sao Paulo)	Prospective randomized controlled trial	30 (15/15)	70.4 (8.2)	POAG (30)	N	GAT	Unstated	800 mL in 5 min	NA	Not Applicable
38	Joanna Przezdziecka-Dołyk 2021	Poland (Wroclaw)	Prospective observational study	49(NA)	Post-XEN: 65 (10) POAG: 68 (8)	Post-XEN (27) POAG (22)	N	GAT, ORA	4 h	10 mL/kg in 5 min	NA	Fluctuation =Peak IOP-baseline IOP Amplitude = Highest IOP– Lowest IOP

GAT Goldmann Applanation Tonometer, PAT Perkins applanation tonometer, NCT non-contact tonometer, ORA ocular response analyzer, XEN XEN Gel Stent, POAG primary open angle glaucoma, OAG open angle glaucoma, PACG primary closed angle glaucoma, XFG pseudoexfoliativeGlaucoma, XFS pseudoexfoliative syndrome, HTG high tension glaucoma, NTG normal tension glaucoma.

The definition of IOP fluctuation varied between studies. Most studies (15 out of 38) defined IOP fluctuation as the difference between the pre-WDT baseline IOP and the peak IOP measured during the WDT. 5 studies [37, 41, 45, 46, 48] defined IOP fluctuation as the percentage of change from the pre-WDT baseline IOP to the WDT peak IOP. Only 1 study [36] defined IOP fluctuation as the difference between the peak IOP and lowest IOP measured during the WDT. 3 studies [32, 37, 38] were included in final meta-analysis – of these, one study [37] defined IOP fluctuation as the difference between the pre-WDT baseline IOP and the peak IOP during the WDT, while the remaining 2 studies [32, 38] did not specify the definition of IOP fluctuation used.

Relationship between WDT results and diurnal IOP curves

9 studies [17, 22, 32, 36–38, 41, 44, 49] with a total of 604 participants examined the relationship between WDT results and diurnal IOP curves. 4 studies [22, 36, 38, 49] (44.4%) included participants with OAG, 2 studies [37, 41] (22.2%) included participants with XFG, 2 studies [17, 32] (22.2%) had a mix of glaucoma subtypes and 1 study [44] (11.1%) included participants with Normal Tension Glaucoma (NTG). The diurnal IOP curve was measured differently among studies, with 7 studies [17, 22, 36–38, 41, 49] (77.8%) collecting IOP results during office hours (between 8am and 6 pm) and 2 studies [32, 44] (22.3%) collecting IOP readings over 24 h.

The largest study, conducted by Vasconcelos-Moraes et al. [22] on 97 participants with primary open angle glaucoma (POAG), found a strong correlation between the peak IOP of the WDT and that of the modified diurnal tension curve (mDTC) (r = 0.780, 95% CI = (0.716, 0.831), p < 0.0001). However, in terms of absolute peak IOP values, only 52.5% of results showed agreement within ±2 mmHg between the IOP peaks of both tests. Notably, Caiado et al. [36] found that the IOP peak of the WDT was higher than that of the mDTC measured while sitting down (diurnal, sitting: 20.0 ± 3.3 mmHg vs WDT:22.5 ± 4.4 mmHg, p = 0.008), but not when mDTC was measured in the supine position (p = 0.207).

2 studies [37, 38] explored the correlation between mean IOPs of the WDT and diurnal IOP curves. Both Olatunji et al. [38] (r = 0.728, p < 0.001) and First et al. [37] (r = 0.884, p < 0.001) found a strong positive correlation between the mean IOPs of the 2 curves.

Meta-analysis

5 studies [22, 32, 37, 38, 49] comprising 310 participants were included in our meta-analysis examining the correlation in peak IOP and IOP fluctuation, between the WDT and the diurnal IOP curve. We excluded the medically-treated group from Medeiros et al. [49] as its CI did not overlap with the mean for our meta-analytic results (Fig. 2a). After exclusion of this study, our results showed a strong positive correlation between the peak IOP from the WDT and the peak IOP of the diurnal IOP curve (r = 0.92, 95% CI= (0.75, 1.08), p < 0.0001) (Fig. 2b). There was however negligible correlation between IOP fluctuation of the WDT and IOP fluctuation of the diurnal IOP curve (r = 0.26, 95% CI= (0.06, 0.47), p < 0.0001) (Fig. 3).

Fig. 2. Forest Plot for correlation of IOP peaks of the WDT and Diurnal IOP curve.

a Forest Plot for correlation of IOP peaks of the WDT and Diurnal IOP curve, before exclusion of medically treated group. b Forest Plot for correlation of IOP peaks of the WDT and Diurnal IOP curve, after exclusion of medically treated group.

Fig. 3.

Forest Plot for correlation of IOP fluctuations of the WDT and Diurnal IOP curve.

There was no significant heterogeneity between studies for both the IOP peak (p = 0.37) and IOP fluctuation (p = 0.64). Publication bias was also negligible based on the Egger’s regression test (p > 0.05), and the funnel plots presented with reasonable symmetry after exclusion of the medically-treated group from Medeiros et al. [49]. (Appendix 4a, 4b, 5).

Of these 5 studies, only 3 studies comprising 189 participants reported on the agreement between the IOP peaks of the WDT and diurnal IOP curve. Meta-analysis revealed that the diurnal curve peak IOP was lower than the WDT peak IOP, however the difference was not significant (mean difference = –2.37 mmHg, 95% LOA = –5.58,0.84, p = 0.147) (Fig. 4).

Fig. 4.

Forest Plot for mean differences between IOP peaks of the WDT and Diurnal IOP curve.

Correlation between WDT and long-term IOP curves

Two studies [20, 46], comprising 85 participants with OAG, investigated the relationship between WDT and long-term IOP curves over a period ranging from 6 months to 4.5 years. Almeida et al. [20] followed participants with stable OAG over 6–12 months, obtaining at least 8 readings, while De Moraes et al. [46] followed participants with POAG over an average period of 4.5 ± 1.3years with at least 5 readings.

Both Almeida [20] (r = 0.52, p < 0.01) and De Moraes [46] (rho=0.76, p < 0.001) observed a moderate to high positive correlation between the peak IOPs of the WDT and long-term IOP curves. Almeida et al. [20] also found a high correlation between the mean IOPs of the 2 methods (r = 0.67, p < 0.01). However, while De Moraes et al. [46] identified a strong positive correlation between the IOP fluctuations of the two curves (rho = 0.82, p < 0.001), Alameda et al. [20] did not find a significant correlation between the fluctuations of the 2 curves (p = 0.45).

Bland Altman plots from the 2 studies reported conflicting results regarding the absolute difference in peak IOP readings between the WDT and long-term IOP curves. Almeida et al. [20] found that long-term IOP had on average a higher IOP peak (mean difference=1.3, 95% LOA= (–7.2,9.7)) and greater IOP fluctuation (mean=3.3, 95% LOA = (–4.9, 11.6)) compared to the WDT, while De Moraes et al. [46] found that long-term IOP produced a lower IOP peak (mean difference = –1.9, 95% LOA = (–5.6, 1.8)) and less IOP fluctuation (mean difference = –9.5, 95% LOA = (–20.9, 1.9)) compared to the WDT. Long-term IOP curves also produced a mean IOP that was on average 0.60 mmHg lower than that of the WDT (p < 0.01) according to Almeida et al. [20].

WDT predicting glaucoma severity

Four studies [16, 23, 35, 39], comprising 282 participants with POAG, investigated the association between WDT results and glaucoma severity. Among these, 3 studies [23, 35, 39] involved participants with asymmetrical VF loss. Glaucoma severity was determined by the VF mean deviation (MD). Three out of four (75%) studies [16, 23, 35] found that WDT results were unable to predict glaucoma severity among participants.

Susanna et al. [23] found that the eye with a worse VF MD exhibited higher peak IOP and IOP fluctuation compared to the better eye (p < 0.001). In contrast, Hatanaka et al. [35] did not find significant differences between the peak IOP (p = 0.570), IOP fluctuation (p = 0.570) and mean IOP (p = 0.851) between the better and worse eye in participants with asymmetrical glaucoma. Similarly, Scoralick et al. [39] reported no significant correlation between VF loss and peak IOP (r = 0.15, p = 0.160), mean IOP (r = 0.15, p = 0.155) and IOP fluctuation (r = 0.17, p = 0.117). Of note, Nicolela et al. [16] found that a later peak IOP correlated with greater VF loss (coefficient = –0.155, 95% CI= (–0.284, –0.025), p = 0.020).

WDT predicting future glaucoma progression

Two studies [34, 48], including 172 participants with POAG, investigated the association between WDT results and future glaucomatous VF progression over 26 months [48] and 28 months [34]. VF progression was established based on the Anderson’s criteria [55], defined by the development of a new defect or worsening of a previous defect. A new defect was defined by the presence of a new cluster of ≥ 3 non-edge points with sensitivities at P < 5% on the pattern deviation plot, with at least 1 of the points at P < 1%. Worsening of a previously damaged region was defined as a deterioration of ≥3 points in that region by at least 10 dB. Both studies found that a higher peak IOP was correlated with an increased likelihood of future glaucoma progression. One of the 2 studies by De Moraes et al. [34] identified peak IOP as a predictor for future progression (Hazard Ratio = 1.11, 95% CI = (1.02, 1.21), p = 0.013), with the Kaplan-Meier curve demonstrating that individuals with a peak IOP of ≥18 mm Hg had a higher chance of progression (p = 0.014).

However, there were conflicting results with regards to the association between IOP fluctuation and VF progression. Susanna et al. [48] reported that participants with greater VF deterioration exhibited a higher IOP fluctuation (34.0 mmHg (3.6) vs 17.2 mmHg (2.8), p < 0.01). In contrast, De Moraes et al. [34] did not find a correlation between IOP fluctuation and VF progression (HR = 1.10; 95% CI = (0.99,1.23), p = 0.074).

Factors affecting WDT results

11 studies [17, 19, 20, 25, 27, 31, 40, 41, 44, 45, 47], including a total of 728 participants, studied the effect of various factors on WDT results.

We classified these factors into 4 categories: (1) patient demographics (including age, BMI, body weight, gender and race) (2) glaucoma-related factors (including type of glaucoma, baseline IOP, type of glaucoma surgery, use of topical medications, number of topical medications, glaucoma severity, visual field parameters, anterior chamber depth, and years of follow up for glaucoma) (3) other ocular factors (including central corneal thickness (CCT), axial length and phakic status), and (4) systemic factors (e.g. the use of hypertensive medications). The outcomes measured in these 11 studies included peak IOP during the WDT, IOP fluctuation during WDT and change in IOP 60 min after the WDT.

The three most studied factors were age, type of glaucoma and CCT.

The effect of age on WDT test results was explored in 4 studies [17, 19, 20, 44]. Only Sakata et al. [44] found a significant association between age and WDT peak IOP (p = 0.013), and between age and the difference between the WDT peak IOP and pre-WDT baseline IOP (p < 0.001). The remaining 3 studies found no significant association between age and IOP fluctuation [19], between age and a > 3 mmHg difference between peak WDT IOP and pre-WDT baseline IOP [17], as well as between age and WDT peak IOP [20].

Differences in WDT results between various subtypes of glaucoma were also compared in 4 studies [17, 19, 27, 40]. 2 studies compared the WDT results of medically-controlled POAG and primary angle closure glaucoma (PACG) and found no significant difference (p < 0.05) between the two groups [19, 40]. Another study, after univariate analysis, found that none of the subtypes of glaucoma studied (primary angle closure disease, NTG, XFG) were significant risk factors for an IOP increase of >3 mmHg during the WDT [17]. Finally, another study showed that neovascular glaucoma experienced the highest difference between peak WDT IOP and pre-WDT baseline IOP (32.2 mmHg), compared to PACG (12.5 mmHg), POAG (18.9 mmHg) and traumatic glaucoma (22.3 mmHg). However, the sample size of participants in this study was small and no p-value was reported [27].

4 studies studied the effect of CCT on WDT results [17, 20, 44, 45]. 2 studies showed that CCT was significantly associated with the WDT peak IOP [20, 44]. However, another study demonstrated that while CCT was associated with an increase in the odds ratio of a WDT IOP response of >3 mmHg, this result was no longer significant after multivariate analysis [17]. The fourth study found no significant correlation between CCT and WDT peak IOP nor IOP fluctuation [45].

WDT in comparing glaucoma treatment outcomes

13 studies [18, 19, 24, 25, 27, 28, 30, 49–53, 56], with a total of 983 participants, compared WDT results following different glaucoma interventions. Of these 13 studies, 4 compared outcomes of different anti-glaucoma medications regimes [24, 50–52], 5 compared outcomes between both medical therapy and surgical therapy [19, 28, 30, 49, 53], and 1 compared outcomes of trabeculectomies versus tube surgeries [25]. Of the remaining 3, one study compared WDT results before and after starting prostaglandin treatment [18], another compared WDT results before and after trabeculectomy [27], and the last compared WDT results between contralateral eyes treated with latanoprost versus selective laser trabeculoplasty [56]

The largest study, by Vetrugno et al. compared WDT results of 280 POAG patients that were treated with different topical medications [24]. There were 7 comparison groups, each comprising 40 patients treated separately with timolol, dorzolamide, brinzolamide, travoprost, latanoprost, bimatoprost, and brimonidine. The findings revealed that medications which enhanced aqueous humor outflow (bimatoprost, latanoprost, travoprost, and brimonidine) exhibited greater IOP stabilization, with lower peak IOPs and shorter durations of IOP elevation compared to medications which decreased aqueous humor production (timolol, dorzolamide and brinzolamide).

Reproducibility of WDT results

3 studies [21, 26, 43], including 143 participants, examined the reproducibility of WDT results. Participants in these studies were medically treated and considered stable. 2 studies [21, 43] repeated the WDT following a 1 or 3–6 month interval, while 1 study [26] repeated the WDT at different times of the day (0700, 1200, 1700).

Overall, the IOP peak of the WDT was found to be more reproducible across repeated tests, compared to IOP fluctuation.

Ozyol et al. [21] demonstrated that the peak IOP of the WDT performed 1 month apart, in both XFG and XFS patients, was highly reproducible (mean difference=0.0 mmHg, 95% LOA = (-3.50,3.40)), with differences within 4 mmHg considered reproducible. Similarly, Medina et al. [26] found that the peak IOP measured by the WDT performed at 3 different timepoints (0700, 1200 and 1700) across 3 occasions spaced 1 week apart were similar, with difference in readings ranging only between 0.10 mmHg to 0.20 mmHg. The largest difference was between 1200 and 1700 (mean difference=0.20 mmHg, 95% LOA:-4.40 to 4.80). A strong correlation was demonstrated among the peak IOPs measured at the 3 timepoints (r > 0.80, p < 0.01), except between 1200 and 1700 (p = 0.362). Babic et al. [43] examined reproducibility of the WDT performed 3–6 months apart and found excellent reproducibility of the IOP peak (0.85 ICC, P < 0.001), with 79.5% of eyes showing a difference of < 2 mmHg between tests.

However, IOP fluctuation was found to be less reproducible. Ozyol et al. [21] found poor reproducibility (defined as a difference of >2 mmHg) in IOP fluctuation (mean difference = –0.10 mmHg, 95% LOA = (–2.30,2.20)) for the WDT performed at different timepoints. Medina et al. [26] also found only weak to moderate correlation between IOP fluctuations when the WDT was performed at 3 timepoints (0700, 1200 and 1700) (r < 0.50), with no significant correlation between 0700 and 1700 (p = 0.148). Larger differences in IOP fluctuations were also observed, ranging from 0.10 to 0.60 mmHg, with the largest difference found between 1200 and 1700 (mean difference = –0.60, 95% LOA = (–6.90,5.70))—similar to that for peak IOP measurements. Finally, Babic et al. [43] also reported poor reproducibility of IOP fluctuation (0.50 ICC, P < 0.001), with higher mean differences between IOP fluctuation readings (mean difference = –0.58, 95% LOA = (-4.98,3.81)) compared to peak IOP measurements (mean difference = −0.47, 95% LOA= (–4.24, 3.30)).

Discussion

Diurnal IOP curves have been shown to be able to predict glaucoma severity and progression as well as predict response to treatment [5, 57, 58]. However, obtaining diurnal IOP curves via multiple IOP measurements in the clinic setting requires significant logistical support, cost and inconvenience to patients. Rebound tonometers such as ICare Home (ICare Finland Oy) and ICare ic100 (ICare Finland Oy, Vantaa, Finland) present useful alternatives, by allowing patients to self-monitor their IOPs at home. However, patients often find these tonometers difficult to use, with IOP results obtained from these tonometers appearing to differ significantly from those obtained using the GAT [59]. Other novel methods have emerged in recent years, including the integration of IOP sensors in intra-ocular lenses [60] as well as contact lenses capable of measuring IOP [61]. While promising, these methods are not yet commercially available and are associated with adverse effects, with rare but potentially sight-threatening complications [62].

Hence, it has been suggested that the WDT may be a more feasible, safer and more convenient surrogate to diurnal IOP curves. Patients with OAG subtypes, such as POAG and XFG, have been demonstrated to have increased resistance in their trabecular outflow meshwork [63], postulated to be due to changes in the quality and amount of the extracellular matrix in the juxtacanalicular region of the trabecular meshwork, leading to impaired aqueous humor drainage [63]. The WDT is understood to stress the trabecular outflow facility of POAG patients, in response to the raised IOP brought about by water ingestion. Another postulated mechanism underlying the WDT involves the raising of the episcleral venous pressure. resulting in a transient period of negative aqueous outflow [64, 65]. In comparison, the WDT has been postulated to cause IOP elevation in PACG patients via different mechanisms. Firstly, the WDT may increase sympathetic stimulation of the iris dilator muscle, resulting in circumferential folding and worsening of the angle closure, precipitating pupillary block and a raised IOP in eyes with narrow angles [66]. Secondly, choroidal expansion may play a role, with a significant increase in choroidal thickness and decrease in anterior chamber depth observed after WDT in ACG patients but not in POAG patients [67]. Water consumption is understood to cause a transient decrease in blood osmolarity and thus pushing fluid from the systemic circulation into the choroidal space down the osmotic gradient. The increase in choroidal volume then pushes upon intraocular compartments, causing an increase in IOP [68, 69]. However, some studies have not been able to determine a definite correlation between choroidal thickness and IOP change [70, 71]. Of note as well, despite differences in structural reactions, angle anatomy and ocular biometrics, PACG and POAG patients have been shown to produce similar peak IOP and IOP fluctuations after the WDT [40].

Our study found a strong correlation in the peak and mean IOPs between both the WDT and diurnal IOP curve, while the correlation in IOP fluctuation between the two tests appeared to be negligible. The strong correlation in peak IOPs between both methods of measurement may be attributed firstly to the WDT’s role as a stress test that evaluates the trabecular outflow facility. OAG patients, with an impaired trabecular meshwork drainage, would not be capable of withstanding the increased stress, expectedly producing both a higher peak IOP as well as mean IOP during the WDT. Lusthaos et al al. [72], using hemoglobin video imaging (HVI), demonstrated that while both glaucomatous and non-glaucomatous patients experienced an increased aqueous outflow in response to the WDT, the increase in outflow was not sustained in glaucoma patients, thus resulting in a persistently elevated IOP following the WDT. Secondly, diurnal IOP variation has also been reported to follow the circadian rhythm with a nocturnal acrophase, postulated to be due to the supine position at night leading to increased episcleral venous pressure [73]. The WDT replicates this increase in episcleral venous pressure in a controlled and shorter timeframe, explaining the correlation between mean and peak IOPs in both tests. That the mean and peak IOPs of the WDT have been successfully shown to be strongly correlated with that of diurnal IOP curves is a notable finding, given that these IOP parameters derived from diurnal IOP curves have been shown to have clinical utility, such as in predicting future glaucoma progression [74].

The negligible correlation between IOP fluctuation in both tests may be related to the baseline IOPs being measured at different timings for both tests, when IOP fluctuation was defined with reference to pre-test baseline IOPs. Most of the WDTs were carried out in the late afternoon (4–5 pm), only after completing the diurnal IOP tests, while the first measurement for most diurnal IOP curves was performed in the morning. IOPs fluctuate throughout the day [4] and different baseline IOPs taken at different times in the day may affect IOP fluctuation outcomes derived from each of the two tests. Of note, Caiado et al. defined IOP fluctuation as the difference between IOP peak and IOP trough, with no reference to baseline pre-test IOPs, and found no significant difference in IOP fluctuations of the WDT and the diurnal IOP curve (p > 0.10) [36].

While our analysis revealed no statistically significant difference in peak IOPs between both the WDT and diurnal IOP curves (mean difference = −2.37, 95% LOA= (–5.58, 0.84), p = 0.147), results trended toward the WDT having a higher peak IOP. This may be consistent with the mechanism of the WDT, where the fluid challenge may more rapidly and significantly increase the episcleral venous pressure within a shorter time compared to the mechanisms underlying the diurnal IOP curve, hence resulting in a higher peak IOP.

With regard to the relationship between WDT results and long-term diurnal IOP curve outcomes, 2 studies demonstrated that the WDT peak IOP may be associated with the peak IOP of long-term IOP measurements [20, 46] However, results for IOP fluctuation appeared to be conflicting, with also no agreement on whether the WDT or long-term diurnal IOP curves yielded a higher peak IOP. Overall, assessment of this relationship may be limited in this review, given the relatively small sample size and differences between both studies in the follow-up duration and type of glaucoma, which may influence results. De Moraes et al. [46] only included 20 POAG patients with a mean follow-up of less than 9 months, while Almeida et al. [20] included more than 60 patients with stable OAG over more than 4 years. Furthermore, De Moraes et al. [46] defined IOP fluctuation as the difference between peak IOP and pre-WDT baseline IOP, while Almeida et al. [20] defined IOP fluctuation differently, as a percentage change from pre-WDT baseline IOP to peak IOP.

WDT results also did not appear to be correlated with current glaucoma severity. While Susanna et al. [23] found that in asymmetrical glaucoma patients the better eye tended to have a lower peak IOP and IOP fluctuation, the differences in peak IOP (0.7 mmHg) and IOP fluctuation (0.8 mmHg) were small. The remaining 3 studies in our review found that the WDT was unable to predict the present extent of VF loss in glaucomatous eyes. The timing of the IOP peak may be a possible metric to predict the severity of current VF loss, but this was only explored by Nicolela et al. [16] with 31 glaucoma patients. However, the WDT showed promise in estimating future glaucoma progression, with a higher peak IOP shown to correlate with greater future VF loss.

Understanding the factors that may influence WDT results is important in evaluating and contextualizing the use of WDT in real-life clinical practice. These factors may also be accounted or controlled for in the design of the WDT protocol for certain patients, or in the analysis of WDT results. A better understanding of the effect of BMI or weight on WDT results for example, may guide adjustment of the amount of water ingested during the WDT, to provide more reproducible and accurate results [47]. Other variables, such as CCT or ocular surface conditions, may also have an impact on IOP measurements [75].

With respect to the reproducibility of WDT results, when the WDT was performed across months and at different timings of the day, the IOP peak measurement remained largely consistent, compared to IOP fluctuation results. Good reproducibility in the context of the WDT is noteworthy, given the inherent variability in individual and ocular metrics, such as body mass and CCT, across periods of time [76, 77]. However, good reproducibility may also be attributed to the relatively short follow-up periods, with the longest follow-up extending up to only 3–6 months, as both systemic and ocular parameters are unlikely to change significantly within a shorter period of time.

The WDT operates on the principle of measuring IOP changes following the administration of a fluid challenge. The aqueous humor outflow capabilities of an eye affect how rapidly it can recover from the IOP elevation caused by the WDT [16, 22]. Therefore, it may be postulated that by measuring the peak IOP of the WDT, the WDT is a reasonable tool to estimate the ability of an intervention to enhance aqueous outflow. This hypothesis may be supported by studies demonstrating that medications that increase aqueous outflow facility (such as prostaglandin analogs) result in lower WDT peak IOPs than medications that decrease aqueous humor production [19, 24].

It remains unclear if WDT results differ between trabeculectomies and tube shunt surgeries. One study found that trabeculectomies and tube shunt surgeries showed a similar pattern of IOP response to the WDT [25], whereas another reported a significantly greater IOP at 60 min after the WDT following tube shunt surgeries, compared to the trabeculectomy group (5.6 ± 3.6 vs. 3.1 ± 4.3; P = 0.007) [19]. However, in comparing medical versus surgical therapy, trabeculectomy surgery resulted in lower IOP elevation [19, 28, 30], a lower peak IOP [28] and less IOP fluctuations [49] than medical therapy. Similarly, deep sclerectomy with collagen implant showed lower average IOP elevation than latanoprost [30], and patients post-XEN Gel Stent implantation demonstrated less IOP fluctuation than those on pharmacological control [53]. These findings are consistent with the expectation of greater IOP-lowering efficacy and lesser IOP fluctuation following surgery compared to medical therapy [78, 79].

Notably, only 4 out of the 38 included studies [17, 19, 27, 40] (10.5%) included patients with PACG, despite PACG comprising roughly 50% of all adult primary glaucoma cases [80]. The ocular response to the WDT appears to differ between POAG and PACG eyes, with PACG eyes possibly experiencing greater choroidal expansion and decrease in anterior chamber depth following the WDT, compared to POAG eyes [67]. The potential role of the WDT in eyes with PACG also differs from that in POAG. Traditionally regarded as a means to assess the outflow facility, severity and future progression in POAG eyes, the WDT has instead been explored as a provocative test to identify patients at a higher risk of developing PACG or experiencing an acute angle closure crisis [66, 81]. However, there have been conflicting findings regarding the WDT’s accuracy as a provocative test for PACG [81], highlighting the need for further investigation and consensus in this area.

This systematic review and meta-analysis possesses inherent limitations. Firstly, differences in study design and methodology were a major source of heterogeneity among studies. For example, methods of obtaining diurnal IOP curves differed among studies, with 7 studies (77.8%) collecting IOP results during office hours (between 8am and 6 pm) while 2 studies (22.3%) collected IOP readings over 24 h. Diurnal IOP peaks may not occur during office hours for all patients [11]. The mDTC was similarly performed at different timings. Of note however, the difference between 24 h and mDTC peak IOPs appears to be modest, demonstrated to be an average of 2.1 mmHg [82]. Furthermore, although mDTCs were performed at various timings in the day, the timings only differed slightly and thus would unlikely affect IOP results significantly. WDT protocols also varied from study to study. Pre-test fasting periods ranged from 2 to 6 h before the WDT, the volume of water drank by participants ranged from 800 ml to 15 ml/kg of body weight and participants were given a range of between 5 and 15 min to consume the water. However, there is no definite evidence suggesting that these differences in protocol influence IOP results [6]. Secondly, participants from different studies were on different glaucoma medication regimes. These differences could not be controlled nor accounted for, and hence may affect WDT and diurnal IOP curve results [25]. Different types of glaucoma within each study might also affect WDT results due to the varying extents of angle-opening and different morphologies of the trabecular meshwork [83]. As most of our studies included participants with OAG, our findings may be less generalizable across other glaucoma subtypes. Fourthly, certain patient populations were excluded from the WDT in our included studies. These included patients with advanced glaucoma, as well as concomitant diseases such as cardiac disease, renal disease, and urinary retention, all of which are contraindicated for the WDT [8]. Our results may therefore also be less applicable to the abovementioned populations. Lastly, a small number of studies (n = 3) and participants (n = 189) were included in our meta-analysis of Bland-Altman plot results, as only a few studies reported on the agreement of IOP peak measurements between the WDT and DC. This might thus affect the robustness of our results for the agreement between the WDT and diurnal curve peak IOPs. Future research in this aspect, with standardized reporting, is recommended.

Conclusion

Nonetheless, to the best knowledge of the reviewers, this review presents the first systematic review and quantitative analysis of available literature regarding multiple aspects of the WDT, which provides a more evidence-based and critical evaluation of the potential role of the WDT in clinical practice. Unlike IOP fluctuation, the peak IOP of the WDT appears to have a stronger correlation with that from both short- and long-term diurnal IOP curves, a greater association with future glaucoma progression and a greater degree of reproducibility. However, varying definitions of IOP fluctuation may have influenced meta-analytic results. Future prospective studies with standardized WDT protocols, as well as consistent definitions and reporting of IOP parameters, may enable more meaningful meta-analysis. Overall, our findings suggest the WDT having a potential role as a more feasible alternative to diurnal IOP measurements, in the monitoring and management of glaucoma in selected clinical contexts.

Author contributions

Methodology, article search and screenings were done by EJ and CXYG. Data Analysis was done by EJ and CPH. Writing of the paper was done by EJ, CXYG and BKB. Review of the manuscript was done by BKB and BCHA.

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41433-024-03107-z.