Article Text

Download PDFPDF

Original article
Therapeutic potential of valproic acid for retinitis pigmentosa
  1. C M Clemson1,
  2. R Tzekov1,
  3. M Krebs2,
  4. J M Checchi1,
  5. C Bigelow3,
  6. S Kaushal1
  1. 1Department of Ophthalmology, University of Massachusetts Medical School, Worcester, Massachusetts, USA
  2. 2Ophthalmology Department, University of Florida, Gainesville, Florida, USA
  3. 3Department of Statistics, University of Massachusetts School of Public Health and Health Sciences, Amherst, Massachusetts, USA
  1. Correspondence to Dr Shalesh Kaushal, Department of Ophthalmology, University of Massachusetts Medical School, 381 Plantation St., Biotech 5, Suite 250, Worcester, MA 01605; shalesh.kaushal{at}


Background/aim To examine the efficacy and safety of valproic acid (VPA) in patients with retinitis pigmentosa (RP).

Methods Thirteen eyes were examined before and after brief treatment (average 4 months) with VPA. Visual fields (VF) for each eye were defined using digitised Goldmann Kinetic Perimetry tracings. VF areas were log-transformed and VF loss/gain relative to baseline was calculated. Visual acuity was measured using a Snellen chart at a distance of 20 feet (6.1 m). Values were converted to the logarithm of the minimum angle of resolution (logMAR) score.

Results Nine eyes had improved VF with treatment, two eyes had decreased VF and two eyes experienced no change, with an overall average increase of 11%. Assuming typical loss in VF area without treatment, this increase in VF was statistically significant (p<0.02). An average decrease (0.172) in the logMAR scores was seen in these 13 eyes, which translates to a positive change in Snellen score of approximately 20/47 to 20/32, which was significant (p<0.02) assuming no loss in acuity without treatment. Side effects were mild and well tolerated.

Conclusion Treatment with VPA is suggestive of a therapeutic benefit to patients with RP. A placebo-controlled clinical trial will be necessary to assess the efficacy and safety of VPA for RP rigorously.

  • Retinitis pigmentosa
  • valproic acid
  • visual fields
  • visual acuity
  • retina
  • apotosis
  • degeneration
  • drugs
  • field of vision

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Retinitis pigmentosa (RP) is a severe neurodegenerative disease of the retina characterised initially by night blindness, with progression to tunnel vision and eventual loss of central vision and total blindness. Targeted therapies for RP are complicated by the identification of more than 40 genes linked to the dominant and recessive forms of this disease. A few new approaches for RP treatment have recently been investigated, including nutritional supplementation, light reduction and gene therapy;1 2 of these, vitamin A supplementation is the most promising,3–5 but its benefits are modest and side effects are problematic. Therefore, there is currently no significant treatment or cure for RP.

Recently we demonstrated the use of retinoids and other small molecules as pharmacological chaperones to increase the yield of properly folded RP mutant rhodopsins in heterologous cell culture.6 We have tested whether other known small molecules can provide similar effects. We identified valproic acid (VPA) through this screen (S. Noorwez and S. Kaushal, unpublished data). Prior work has suggested that VPA is a potent inhibitor of histone deacetylase (HDAC)7 and the inflammatory response pathway via apoptosis of microglial cells.8–10 In addition, VPA is known to downregulate complement proteins11 and increase the levels of various neurotrophic factors.12 Thus, VPA has a unique biological profile suitable for treating retinal diseases. VPA was approved by the Food and Drug Administration for use as a broad-spectrum anticonvulsant in 1978 and is also approved for acute and maintenance therapy of mania in bipolar disease and for migraine prophylaxis. VPA and its derivative, divalproex sodium, is used off-label for a variety of indications including chronic pain syndromes, cancer therapy and schizophrenia.13 Collectively, this body of evidence suggests that VPA may be an appropriate therapy for patients with retinal dystrophies.

Here, we present the results of a pilot analysis of the effect of short-term VPA treatment on vision function of patients with RP. The results presented here will be used to inform a larger clinical trial to test the safety and efficacy of VPA for RP.

Patients and methods

This study is a retrospective chart review of patients with RP who were treated off label with VPA at the University of Florida Ophthalmology Department clinic between December 2007 and January 2009. The data reviewed from these patients records were not collected originally with the intent to study. Fourteen RP patients were identified; of these seven had adequate baseline and follow-up visual fields (VF) (the VF from one eye of subject 5 were excluded due to poor quality). The length of treatment included in this analysis was based on the data available in the record and varied from 2 to 6 months. The dosage of VPA varied from 500 to 750 mg/day, which is lower than the dosage typically used for anticonvulsant therapy. A daily dosage of 500 mg was originally chosen as it is approximately half the dosage prescribed for other indications; as this was well tolerated, several patients were prescribed 750 mg. Patient demographics, diagnosis, family history, genotype, best corrected visual acuity (BCVA—which was converted to the logarithm of minimal angle of resolution (logMAR)), dosage of VPA, length of treatment, blood chemistries including alanine aminotransferase (ALT), aspartate aminotransferase (AST), ammonia, and electrolyte and blood cell panels including Na, K, Cl, bicarbonate, creatinine, white blood cells with differential, red blood cells and platelets) and reported side effects were all recorded.

For each patient, intact baseline and follow-up visual field areas were defined using the existing Goldmann Kinetic Perimetry tracings (isopter V4e) from each eye. The tracings were digitised and the corresponding areas of functioning retina (in mm2) were calculated based on the method used by Dagnelie.14

We defined the change in visual field (mm2) as a simple measure of per cent change from baseline:

Per cent change from baseline=100×((follow-up in mm2baseline in mm2)/baseline in mm2)

Improvement in VF was defined as >2% increase, a loss in VF area was defined as >2% decrease, while VF was considered unchanged if the follow-up value was within 2% of baseline.

VF loss in RP does not occur at a linear rate, and Massoff et al demonstrated that it may decline exponentially, with an estimated loss of about 0.10 loge units per year.15

VF areas (mm2) were loge-transformed (logVF) and the difference between follow-up and baseline was calculated (ΔlogVF). To calculate average per cent change in VF over the course of treatment, the average difference across all eyes was calculated and converted to area (mm2). This value was used as the follow-up value in the above formula, and the baseline value used was 314 mm2 (which was the average baseline from all subjects).

Using varying estimates from the literature regarding the natural history of VF loss in RP patients,15–17 we hypothesised that patients without treatment would lose either 0.0, 0.011 or 0.033 logVF over the average length of treatment (4 months). Data were not assumed to be distributed normally and significance levels were calculated using the Wilcoxon signed rank test. Statistical analysis was performed using Graph Pad Prism (La Jolla, California, USA).

Visual acuity was measured using a Snellen chart at a distance of 20 feet (6.1 m). Values were converted to the logarithm of the minimum angle of resolution (logMAR) score for statistical analysis. Significance levels were calculated using the Wilcoxon signed rank test assuming that visual acuity would not change without treatment.


Table 1 summarises the average characteristics of the RP patients included in this analysis. Patients' ages ranged from 16 to 56 (mean 36) years, with five men included. Most patients (n=6, 86%) had a family history or reported genotyping suggestive of an autosomal dominant form of RP (ADRP). The length of treatment on VPA was short, with a range from 2 to 6 months. Overall, average visual acuity was 20/47 per eye (range 20/20 to count fingers at 3 feet (0.9 m) or 20/4000).

Table 1

Study sample characteristics

Analysis of change in functioning retina

VF were measured using kinetic perimetry; tracings were digitised and the areas determined by isopter V were converted into areas of functioning retina. Examples of baseline and follow-up perimeter tracings are shown in figure 1A. The patient whose VF is shown on figure 1A (patient 6, table 1) had two follow-ups within a short period of time (4 weeks) and VF results were stable, suggesting stability in vision function gain as a result of treatment. Line graphs of baseline and follow-up values of VF areas for all seven study subjects, for each eye and the average values per patient, are shown in figure 1B. Baseline and follow-up data are plotted according to their time of measurement, thus accounting for duration of treatment; these graphs depict the individual slopes of change in VF. Using a difference from baseline of ±2% as a criterion for change, nine eyes (five right eyes and four left eyes) had improved intact VF after brief treatment with VPA, while two eyes lost VF area (one right eye and one left eye) and two eyes (one right and one left eye) had no change in VF. Table 2 lists the values for per cent change from baseline and difference in logVF as described in the Methods.

Figure 1

Visual field after treatment with valproic acid (VPA). (A) Goldmann kinetic perimetry tracings from patient 6 at baseline (left) and after 6 months of VPA treatment (middle – red), overlap of baseline and follow-up (right). (B) Change in visual field: baseline and follow-up visual field areas are plotted for each individual patient (length of follow-up varied for each patient). Areas were analysed by right eye, left eye and average of both eyes. Values correspond to the individual subjects.

Table 2

Per cent change in visual field

Some studies estimated an average visual field loss of 0.10 loge units per year (equivalent to ∼10.5%/year) or about 0.033 loge units (∼3.52%) for 4 months,15 which is the average length of VPA treatment in this study. The change in logVF (ΔlogVF) from baseline is presented in table 2 and figure 2. The average change for all 13 eyes over the course of treatment was +0.164±0.298 (range −0.012 to +0.942) loge mm2, corresponding to an average increase of about 35 mm2. Assuming a baseline area of 314 mm2 (the average baseline of subjects in this study), this translates to an 11% increase in area of functioning retina.

Figure 2

Change in visual field. Goldmann kinetic perimetry tracings (isopter V4e) from each eye were digitised and areas (mm2) were calculated as described in Methods and log-transformed. Scatterplots of change in log-transformed visual field (VF) over the course of treatment are shown for left eyes, right eyes and all eyes combined. Mean value is shown by thin bar and standard deviation is represented by upper and lower dark bars. VPA, valproic acid.

We performed an exploratory assessment to see if this effect was significant. We compared the change in VF for eyes on VPA to two theoretical rates of vision loss with no treatment. The first assumption was that there would be no loss of VF without treatment, while the second assumption was that eyes would lose an average of 0.033 loge units or 3.5%.15 The third assumption was for a more conservative estimate of visual field loss based on Berson et al.17 Significance levels were calculated using the Wilcoxon signed rank test (table 3). A significant difference (p<0.006) in VF loss was seen for eyes on VPA relative to the natural history of the disease.15 Even if one assumes a slower rate of decline in visual field loss of 0.011 logVF unit or 1.5% for 4 months duration of VPA treatment (based on data from Berson et al17, the difference is statistically significant (p<0.02).

Table 3

Statistical analysis of changes in VF with VPA relative to typical loss in RP patients

Visual acuity

An overall increase in visual acuity on VPA treatment was observed (table 1). A decrease in logMAR units from baseline to follow-up is indicative of improved BCVA. When analysed by eye, average logMAR for all right eyes decreased at baseline from 0.457±0.515 to 0.260±0.380 at follow-up while average logMAR values for the left eyes decreased from 0.300±0.332 to 0.153±0.113. The average change in logMAR across all eyes was a loss of 0.172±0.269 (range −0.824 to 0) log units, which translates to a positive change in Snellen score of approximately 20/47 to 20/32. Assuming no loss in visual acuity without treatment this change in acuity was significant (p<0.02).

Assessment of possible harm

We performed a conservative analysis to explore the possibility of any potential negative effects of VPA on patient's visual field. For purposes of this analysis, an event of negative effect of VPA was indicated by net loss in visual field from baseline greater than 2%. This is a conservative approach because, without any treatment, the natural progression of RP includes a potential for significant short-term deterioration in visual field.15–17 Of the 13 eyes examined, two experienced worsening of their visual field (figure 1). No abnormal liver function or blood chemistries were noted in the study sample. The most common side effects were mild and included tiredness (10%) and stomach irritation (13%).


RP is a blinding disease with no robust treatment options. The visual field areas of five of seven RP patients increased with a short treatment of VPA. Encouragingly, in one case (patient 6), the significant improvement in functioning retinal area was confirmed at two time points (23 and 27 weeks). While visual acuity is not always a reliable outcome measure for RP given that photoreceptor degeneration typically begins in the periphery and progresses to the central macula in only the latest stages of disease, we observed an overall improvement in acuity while being treated with VPA. These positive results are encouraging given that the VPA dose used in this study was about 60% lower than the typical dose for epilepsy or the dose used in a recently published clinical trial for amyotrophic lateral sclerosis.18

While prior accounts of limited delay of progression of photoreceptor loss in RP patients have been reported with nutritional supplementation such as vitamin A3–5 or treatments such as hyperbaric oxygen therapy,19 to the best of our knowledge, this is the first reported case of improvement of vision function in patients with RP as a result of pharmacological treatment.

VPA is widely used as an anticonvulsant and mood stabiliser and its efficacy in these capacities is probably mediated via its ability to affect gamma-aminobutyric acid (GABA) levels through glutamic acid decarboxylase and GABA transaminase modulation.20 21 It is interesting to speculate on how VPA may act as a retinal therapeutic. We first considered VPA for RP as it was identified using our heterologous cell culture screen for small molecules that can increase the yield of properly folded RP mutant rhodopsins (S Noorwez and S Kaushal, unpublished data). Recent evidence suggests that VPA may work at the level of cell death protection or inflammatory mediation as its neuroprotective properties have been well documented,12 22 23 and it can downregulate the photoreceptor-specific inflammatory response pathway via apoptosis of microglial cells.8–10 Furthermore, VPA is known to be a potent inhibitor of HDAC.7 23 24 A particularly exciting property of VPA has recently been documented that suggests that it has the unique ability to reverse photoreceptor damage: VPA can induce cells to differentiate in culture.7 Moreover, VPA has been shown to stimulate glial cells to differentiate into photoreceptor-like cells.25

While the results presented here are promising, this analysis has several limitations. Only seven patients were analysed and the length of follow-up was brief (an average of 4 months). Patients in this analysis were not thoroughly genetically characterised, and it is possible that genetic variation in known RP genes might account for the variability in therapeutic response to VPA.

In summary, VPA offers an exciting new potential therapy for RP, a tragic blinding disease with no good treatment options. The results of our preliminary clinical analysis in conjunction with the prior in vitro data suggest that VPA may be an effective treatment for photoreceptor loss associated with RP. We plan to use this study as the basis for a placebo-controlled clinical trial with patients with well characterised RP genotypes to fully evaluate the efficacy and safety of VPA as a treatment for RP.


The authors would like to thank Linda Stein for her efforts in editing this publication.


View Abstract


  • Linked article 175364.

  • Funding Vision Research Fund at the University of Massachusetts Medical School.

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the University of Massachusetts Medical School Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • PostScript
    Mary J van Schooneveld L Ingeborgh van den Born Mies van Genderen Jan-Geert Bollemeijer
  • PostScript
    Christine M Clemson Radouil Tzekov Jenna M Checchi Mark Krebs Shalesh Kaushal