Effect of Sleeping in a Head-Up Position on Intraocular Pressure in Patients with Glaucoma

Yvonne M. Buys, MD, FRCSC,1 Tariq Alasbali, MD,1,2 Ya-Ping Jin, PhD,1,3 Michael Smith, MBChB, FRCOphth,1 Pieter Gouws, MB,1,4 Noa Geffen, MD,1 John G. Flanagan, BSc, PhD, FAAO,1,5 Colin M. Shapiro, BSc, MBBS, PhD,6 Graham E. Trope, MB, PhD, FRCSC1

Purpose: To determine whether a 30-degree head-up sleeping position decreases nocturnal intraocular pressure (IOP) compared with lying flat in patients with glaucoma. Design: Prospective, nonrandomized comparative case series. Participants: Seventeen eyes of 17 patients with glaucoma with controlled IOP and new disc hemorrhage. Methods: Patients with a new disc hemorrhage despite well-controlled IOP were evaluated in a sleep laboratory on 2 separate nights, the first night lying flat and the second night in a 30-degree head-up position. Intraocular pressure and blood pressure (BP) were measured every 2 hours from 6 PM to 8 AM. For the 6 PM,8PM,10PM, and 8 AM measurements (awake period) the subjects were sitting for both nights. For the midnight, 2 AM,4AM, and 6 AM measurements (sleep period), the subjects were supine for the first night and 30 degrees head up for the second night. Main Outcome Measure: Difference in IOP during the sleep period (midnight to 6 AM) between lying flat and 30-degree head-up positions. Results: Seventeen eyes of 17 patients were included. There were no significant differences (Pϭ0.68) between the 2 study visits in IOP during the awake period (6 PM,8PM,10PM, and 8 AM) when patients were sitting upright. During the sleep period (midnight to 6 AM) the mean IOP was 3.2 mmHg lower in the 30-degree head-up position compared with the flat position (Pϭ0.03; 95% confidence interval, 0.25–6.1 mmHg). Sixteen of 17 patients (94.1%) had lower IOP in the 30-degree head-up position. The reduction in IOP in the 30-degree head-up position was 20% or more in 35% of patients (6/17). There were no differences in BP or ocular perfusion pressure between the 2 positions. Conclusions: The 30-degree head-up sleeping position lowers IOP compared with the flat position. Although this effect varies between individual patients, mean IOP was 20% lower in one third of patients in this series. Financial Disclosure(s): The authors have no proprietary or commercial interest in any materials discussed in this article. Ophthalmology 2010;117:1348–1351 © 2010 by the American Academy of Ophthalmology.

It has been reported that intraocular pressure (IOP) in- with a new disc hemorrhage were considered for inclusion in the creases when lying flat by as much as 6 mmHg in both study. Exclusion criteria were corneal disease and inability to lie normal patients and patients with glaucoma.1 This supine flat. Written informed consent was obtained from each subject IOP increase1–3 may play a role in progression of glaucoma. before participating in the study. A recent study suggested that IOP should be measured in All subjects were admitted to the Sleep Unit of the University both the sitting and supine positions to predict diurnal Health Network at Toronto Western Hospital from 6 PM to 8 AM on pressure fluctuations.4 Supine increases in IOP are likely due to 2 nonconsecutive nights. From 6 PM to 10 PM, the subjects were awake and encouraged to assume their normal indoor activities and increased episcleral venous pressure.5,6 If patients with glau- were walking, standing, or sitting. At 10 PM, the subjects were left coma have a supine IOP elevation, it is possible that the in a quiet, darkened room and encouraged to sleep. The subjects head-up position during sleep may reduce this postural re- slept in a flat position on the first night and with their head elevated sponse. Because it is not reasonable to expect patients to sleep 30 degrees using a wedge pillow on the second night (Fig 1). At 8 in the 90-degree sitting position, we designed this study to AM the subjects were awakened. Intraocular pressure was measured determine whether sleeping at a 30-degree head-up angle can using a Tonopen (Reichert, Inc., Depew, NY) in the eye with the blunt the postural IOP response seen during the nocturnal disc hemorrhage. For each time point, IOP was recorded as the supine position in patients with glaucoma. average of 2 consecutive measurements that were within 2 mmHg and both with Ͻ5% error as indicated on the Tonopen. Brachial blood pressure (BP) was measured using an automated sphygmo- Materials and Methods manometer. The IOP and BP were measured every 2 hours. Mea- surements were made in the sitting position at 6 PM,8PM,10PM, This study was approved by the Research Ethics Board of the and 8 AM (awake period) each night. During the sleep period University Health Network at the University of Toronto and con- (midnight, 2 AM,4AM,6AM) the subjects were gently awakened ducted in accordance to the Declaration of Helsinki. Patients with and the IOP was measured in the supine position, flat, or 30 treated glaucoma (Table 1) and IOPs in the teens who presented degrees head up depending on the study visit.

1348 © 2010 by the American Academy of Ophthalmology ISSN 0161-6420/10/$–see front matter Published by Elsevier Inc. doi:10.1016/j.ophtha.2009.11.015 Buys et al ⅐ Effect of Sleeping in a Head-Up Position on Intraocular Pressure

The main outcome was IOP from midnight to 6 AM (sleep period) in the flat versus 30-degree head-up position. Secondary outcomes were IOP at all times, systolic BP (sBP), diastolic BP (dBP), mean BP (mBP), and ocular perfusion pressure (OPP). Mean BP was calculated as the dBP plus one third the difference between sBP and dBP. The OPP was calculated in the sitting and supine position using the equations previously published by Liu et al.7 For the 30-degree head-up position, Dr. Anders Bill, who recommended the equations used by Liu et al,7 was contacted and provided the equation. The formulas for calculating OPP were as follows: Sitting OPPϭ0.68ϫmBPϪIOP Supine OPPϭ0.88ϫmBPϪIOP 30 degrees head up OPPϭ0.84ϫmBPϪIOP. Data were analyzed using profile analysis with unstructured covariance to take into account the correlation of repeated readings from the same individuals. A 2-sided P value Ͻ0.05 was consid- ered statistically significant. If the 95% confidence interval of the difference between the 2 sleeping positions did not include 0, the difference was considered statistically significant as well. Analyses were performed with the Proc Mixed in SAS 9.1 (SAS Inc., Cary, NC). A sample size calculation determined that 17 subjects would be required to detect a difference of 2 mmHg at a standard deviation of 2.8 mmHg with a power of 80%.

Results

Seventeen patients completed the 2 nights. The subjects’ demo- graphics are summarized in Table 1. There were 9 female and 8 male patients with a mean age of 65.4Ϯ11.1 years (range, 48.0– 86.3 years). Thirteen of the subjects had primary-open angle glaucoma or normal tension glaucoma. The disc hemorrhage was in the right eye in 6 cases and the left eye in 11 cases. Patients were using a mean of 2.6 glaucoma medications (range, 1–4). Four Figure 1. Sleep positions: Flat position on first night (A) and 30-degree patients were receiving oral antihypertensive therapy. head-up position on a separate night (B). This position was achieved by Figure 2 shows the IOP at each time point for each night. There using a wedge pillow. were no significant differences at any time between the 2 nights

during the awake period when patients were sitting. The IOP was Table 1. Patient Demographics significantly lower in the 30-degree head-up position compared with the flat position at 2 AM,4AM, and 6 AM. The mean IOP during Variable Result the sleep period (midnight to 6 AM) was 15.4Ϯ4.4 mmHg in the 30-degree head-up position compared with 18.6Ϯ4.0 mmHg in the Mean age (range) 65.4Ϯ11.1 yrs (48.0–86.3 yrs) flat position (Pϭ0.03). During the sleep period, the IOP was on Gender 9 F; 8 M Diagnosis 7 NTG average 3.18 mmHg lower in the 30-degree head-up position 6 POAG 1 pigmentary glaucoma 1 angle recession glaucoma 1 ACG 1 uveitic glaucoma Eye with disc hemorrhage 6 OD; 11 OS Mean No. of glaucoma medications 2.6 (1–4) (range) Previous laser trabeculoplasty 7 Previous trabeculectomy 5 Hypertension 4 Coronary artery disease 2

ACG ϭ angle closure glaucoma; F ϭ female; M ϭ male; NTG ϭ normal Figure 2. Mean IOP from 6 PM to 8 AM with both groups supine during tension glaucoma; OD ϭ right eye; OS ϭ left eye; POAG ϭ primary the highlighted sleep period. At all other times both groups were sitting. open-angle glaucoma. IOP ϭ intraocular pressure.

1349 Ophthalmology Volume 117, Number 7, July 2010

Table 2. Awake and Sleeping Systolic, Diastolic, and Mean Blood Pressure and Ocular Perfusion Pressure for the Two Nights*

Awake Period Sleep Period Variable (mmHg) (6 PM to 10 PM;8AM) (midnight to 6 AM) sBP Flat 136.1Ϯ21.1 127.4Ϯ22.6 30 degrees head up 130.5Ϯ16.3 121.4Ϯ20.2 P value 0.39 0.42 dBP Flat 80.6Ϯ9.6 72.3Ϯ10.5 Ϯ Ϯ Figure 3. Mean IOP during the nocturnal period for each subject in flat 30 degrees head up 76.1 6.2 72.7 9.8 and 30-degree head-up positions. IOP ϭ intraocular pressure. P value 0.12 0.91 mBP Flat 99.1Ϯ12.7 90.6Ϯ13.6 compared with the flat position (95% confidence interval, 0.25–6.1 30 degrees head up 94.2Ϯ7.9 88.9Ϯ12.9 mmHg). Sixteen of 17 patients (94.1%) had lower IOP in the P value 0.19 0.70 30-degree head-up position (Fig 3). The reduction in IOP at the OPP 30-degree head-up position was 20% or more in 35% of patients Flat 53.6Ϯ6.3 61.6Ϯ11.2 Ϯ Ϯ (Fig 4). The range in IOP from 6 PM to 8 AM was 10.0Ϯ4.3 mmHg 30 degrees head up 49.8 4.1 59.3 9.8 for the supine position and 6.9Ϯ3.1 for the 30-degree head-up P value 0.05 0.52 position (Pϭ0.02) Table 2 summarizes the results of BP and OPP during the awake dBP ϭ diastolic blood pressure; mBP ϭ mean blood pressure; OPP ϭ and sleep periods. There was no significant difference in sBP, dBP, ocular perfusion pressure; sBP ϭ systolic blood pressure. mBP, and OPP between the flat and 30-degree head-up positions. *During the awake period, both groups were sitting. During the sleep period, groups were in the flat position 1 night and the 30-degree head-up position 1 night. Discussion Studies have shown that consistent and significant low- We hypothesized that patients with progressive glaucoma ering of IOP in patients with glaucoma can slow or prevent despite low IOP during office hours may have a postural IOP progression.10 However, some patients with glaucoma con- response contributing to progression. Because disc hemor- tinue to progress despite low IOP measured during office 8,9 rhages have been associated with glaucoma progression, we hours.17 A 6 mmHg increase in IOP on lying down in decided to investigate the postural IOP response in patients normal patients as well as in patients with glaucoma is a with glaucoma with a new disc hemorrhage and IOP in the well-described phenomenon.1 This effect may play a role in teens. We thought that the relatively simple maneuver of the progression of glaucoma.1–3 Krieglstein et al6 suggested raising the head of the bed when sleeping may significantly that the IOP increase is correlated with changes in episcleral decrease the postural IOP response and possibly improve dis- venous pressure and ophthalmic arterial pressure. Alter- ease control. We therefore investigated whether sleeping at a ations in the rate of uveoscleral outflow may also have an 30-degree head-up position improves nocturnal IOP control. influence on the IOP. We found a significant reduction in IOP and range in IOP in Among the numerous factors that can affect diurnal the 30-degree head-up position compared with the flat position. variation in IOP, body posture is highly significant and There was no effect on BP or OPP. To our knowledge, this somewhat modifiable.11,12 Medical treatment lowers IOP in article is the first to report that sleeping in a 30-degree head-up both the seated and supine positions but does not alter the position decreases IOP in patients with glaucoma. response of IOP to postural change.13 It is known that timolol has little IOP-lowering effect at night.14,15 Although prostaglandin analogs, dorzolamide, and brimonidine lower IOP during the nocturnal period, the nocturnal effect seems to be less than the effect during the diurnal period.15,16 It has also been reported that the postural response is unaffected by trabeculectomy17 and argon laser trabeculoplasty.18 To date, no one has offered a technique to reduce or prevent IOP elevation induced by postural change. The head-up sleeping position effectively achieves this in a substantial number of patients, although not all.

Study Limitations

Figure 4. Percentage decrease in IOP for 30-degree head-up position There were several limitations in our study. First, the study compared with flat position for each subject. IOP ϭ intraocular pressure. includes only 17 patients with normal office hours IOP. As

1350 Buys et al ⅐ Effect of Sleeping in a Head-Up Position on Intraocular Pressure a result, we do not know whether normal individuals or Albrecht Von Graefes Arch Klin Exp Ophthalmol 1978;206: individuals with elevated IOP have a similar or different 99–106. response to the head-up sleeping position. Further work is 7. Liu JH, Gokhale PA, Loving RT, et al. Laboratory assessment needed to determine this. Second, we were not able to of diurnal and nocturnal ocular perfusion pressures in humans. monitor intracranial pressure change, another factor that J Ocul Pharmacol Ther 2003;19:291–7. 8. Drance S, Anderson DR, Schulzer M, Collaborative Normal- might change in the head-up position and influence trans- 19 Tension Glaucoma Study Group. Risk factors for progression laminar pressure gradients. Finally, our calculations of of visual field abnormalities in normal-tension glaucoma. OPP are based on theoretic formulas that may not reflect the Am J Ophthalmol 2001;131:699–708. real physiologic status of ocular perfusion. 9. Leske MC, Heijl A, Hussein M, et al, Early Manifest Glau- In conclusion, the 30-degree head-up sleeping position coma Trial Group. Factors for glaucoma progression and the lowers nocturnal IOP compared with the flat position. Al- effect of treatment: the Early Manifest Glaucoma Trial. Arch though this benefit varies between individual patients, the Ophthalmol 2003;121:48–56. mean IOP was 20% lower in one third of patients in this 10. AGIS Investigators. The Advanced Glaucoma Intervention series. Further research is needed to determine whether the Study (AGIS): 7. The relationship between control of intraoc- 30-degree head-up position during sleep can slow or stop ular pressure and visual field deterioration. Am J Ophthalmol progressive disc damage. The results of this study suggest 2000;130:429–40. that in the management of glaucoma one should not ignore 11. Kothe AC. The effect of posture on the intraocular pressure and the pulsatile ocular blood flow in normal and glaucoma- nighttime pressure increases that occur when sleeping in the tous eyes. Surv Ophthalmol 1994;38(suppl):S191–7. flat position. Clinicians should consider recommending sim- 12. Krieglstein G, Langham ME. Influence of body position on the ple postural changes to patients with features of progressive intraocular pressure of normal and glaucomatous eyes. Oph- glaucoma despite good IOP during office hours; however, thalmologica 1975;171:132–45. further studies are still required to prove whether such 13. Smith DA, Trope GE. Effect of a beta-blocker on altered body postural changes help prevent progression. position: induced ocular hypertension. Br J Ophthalmol 1990; 74:605–6. 14. Orzalesi N, Rossetti L, Invernizzi T, et al. Effect of timolol, References latanoprost, and dorzolamide on circadian IOP in glaucoma or ocular hypertension. Invest Ophthalmol Vis Sci 2000;41: 1. Liu JH, Bouligny RP, Kripke DF, Weinreb RN. Nocturnal 2566–73. elevation of intraocular pressure is detectable in the sitting 15. Liu JH, Kripke DF, Weinreb RN. Comparison of the nocturnal position. Invest Ophthalmol Vis Sci 2003;44:4439–42. effects of once-daily timolol and latanoprost on intraocular 2. Noël C, Kabo AM, Romanet JP, et al. Twenty-four-hour time course pressure. Am J Ophthalmol 2004;138:389–95. of intraocular pressure in healthy and glaucomatous Afri- 16. Orzalesi N, Rossetti L, Bottoli A, et al. The effect of latano- cans: relation to sleep patterns. Ophthalmology 2001;108:139–44. prost, brimonidine, and a fixed combination of timolol and 3. Liu JH, Zhang X, Kripke DF, et al. Twenty-four-hour intraoc- dorzolamide on circadian intraocular pressure in patients with ular pressure pattern associated with early glaucomatous glaucoma or ocular hypertension. Arch Ophthalmol 2003;121: changes. Invest Ophthalmol Vis Sci 2003;44:1586–90. 453–7. 4. Fogagnolo P, Orzalesi N, Ferreras A, Rossetti L. The circadian 17. Parsley J, Powell RG, Keightley SJ, Elkington AR. Postural curve of intraocular pressure: can we estimate its characteristics response of intraocular pressure in chronic open-angle glaucoma during office hours? Invest Ophthalmol Vis Sci 2009;50: following trabeculectomy. Br J Ophthalmol 1987;71:494–6. 2209–15. 18. Singh M, Kaur B. Postural behavior of intraocular pressure 5. Tarkkanen A, Leikola J. Postural variations of the intraocular following trabeculoplasty. Int Ophthalmol 1992;16:163–6. pressure as measured with the Mackay-Marg tonometer. Acta 19. Berdahl JP, Fautsch MP, Stinnett SS, Allingham RR. Intra- Ophthalmol (Copenh) 1967;45:569–75. cranial pressure in primary open angle glaucoma, normal 6. Krieglstein GK, Waller WK, Leydhecker W. The vascular tension glaucoma, and ocular hypertension: a case-control basis of the positional influence of the intraocular pressure. study. Invest Ophthalmol Vis Sci 2008;49:5412–8.

Footnotes and Financial Disclosures

Originally received: July 17, 2009. 6 Department of Cell and Systems Biology, University of Toronto, To- Final revision: October 17, 2009. ronto, Ontario, Canada. Accepted: November 9, 2009. Presented at: The Association for Research in Vision and Ophthalmology, Available online: February 25, 2010. Manuscript no. 2009-959. May 7, 2009, Fort Lauderdale, Florida. The Canadian Ophthalmological 1 Department of Ophthalmology and Vision Sciences, University of To- Society Meeting, June 22, 2009, Toronto, Ontario, Canada. The World ronto, Toronto, Ontario, Canada. Glaucoma Congress, July 8–11, 2009, Boston, Massachusetts. 2 Department of Ophthalmology, King Faisal University, Dammam, Saudi Arabia. Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials 3 Dalla Lana School of Public Health, University of Toronto, Toronto, discussed in this article. Ontario, Canada. Funding: Department of Ophthalmology and Vision Sciences Practice Plan. 4 Department of Ophthalmology, Conquest Hospital, Hastings, United Correspondence: Kingdom. Yvonne M. Buys, MD, FRCSC, Toronto Western Hospital, 399 Bathurst 5 School of Optometry, University of Waterloo, Waterloo, Ontario, St., EW6-405, Toronto, Ontario, Canada, M5T 2S8. E-mail: y.buys@ Canada. utoronto.ca.

1351