INTRODUCTION  

Central  corneal  thickness  plays  a  major role in the management of many types of glaucoma.  Ocular  hypertensive  (OHT) studies  (Kass  et  al.  2002,  Gordon  et  al. 2002)  has  not  only  reported  thin  cornea as  a  predictor  for  the  development  of primary  open  angle  glaucoma  (POAG) but Copt et al (1999) has also reclassified more  than  56%  of  OHT  patients  based on  the  corrected  central  corneal thickness  (CCT).  Additionally,  Kim  and Chen  (2004)  demonstrated  progression of  disease  was  more  likely  in  POAG patients  with  thinner  corneas,  hence giving  another  insight  in  the  role  of  CCT in  managing  glaucoma  patients. Correlation  between  corneal  thickness and  the  severity  of  the  disease  in  the POAG  group  has  been  described  by many  authors  using  advanced  glaucoma intervention study (AGIS) scoring (Kim & Chen  2004,  Herndon  et  al.  2004).  However,  this  relationship  is  not  well described for the NTG group of patients. 

Optical  coherence  tomography  shows good  topographic  relationship  between structural  damage  and  functional  loss (visual  field  loss)  (Wollstein  et  al.  2004, Wollstein et al. 2005). The average RNFL thickness  was  found  to  have  the  strong- est  correlation  with  severity  of  the  dis- ease,  followed  by  the  inferior  and  supe- rior  quadrant  measurements. The  aim  of this  study  was  to  investigate  a  possible correlation  between  the  severity  of  glau- coma  (based  on  RNFL  thickness  as measured  by  OCT)  and  CCT  among NTG and POAG group of patients.

MATERIALS AND METHODS

Patients 

An  observational  study  of  patients  who were treated for primary open angle and normotensive  glaucoma  was  carried  out at  UKMMC,  Kuala  Lumpur  from  January 2006  to  April  2007.  This  study  was  ap- proved  by  the  Medical  Research  and Ethics Committee UKM.   

Design 

Medical  records  of  the  patients  were  re- viewed  retrospectively  for  selection  of cases. Sample size was calculated using the sample size calculation program ver- sion  2.1.30  February  2003.  High  tension glaucoma  was  diagnosed  if  the  baseline IOP  prior  to  treatment  was  more  than 21mmHg  with  the  presence  of  open  an- gle  on  gonioscopy  and  glaucomatous optic  disc  changes.  Those  with  baseline readings less than 21mmHg were classi- fied  as  having  normal  tension  glaucoma (NTG).  All  patients  had  glaucomatous visual  field  defects  as  confirmed  by  the Humphrey  Visual  Field  test  and  only  pa- tients  who  were  on  medical  treatment with  no  previous  surgery  were  recruited for  this  study.  Patients  with  significant anterior  segment  disease  causing  poor signal  strength  on  the  OCT  were  ex- cluded  from  the  study.  Other  exclusion criteria  were  those  with  diabetic  retino- pathy, contact lens user and high myopia of more than  -  6 dioptre. The  worse eye was chosen for the study. 

Measurement of CCT was done using a specular  microscope  SP-3000P  (Topcon Corporation, Tokyo). This method used a non  contact  technique  to  measure  the corneal thickness, which was reported to have  better  repeatability  compared  to ultrasound  pachymetry  (Bovelle  et al. 1999).  Three  readings  with  a standard deviation  of  less  than  3µm  were  used, and  the  average  was  calculated  for  the analysis.  Subsequently,  measurement  of RNFL was done by a single operator us- ing  the Stratus  3 OCT  (Carl  Zeiss  Medi- tech,  Dublin,CA)  on well-dilated  eyes. Good  signal  strength  of  7  or  more  was included  for  analysis.  Diagnosis  was confirmed  by  looking  at their  previous record  of  baseline  IOP  before  treatment and patients who were treated elsewhere with  no known  pretreatment  IOP  were excluded  from  the study.  Purpose  and details of the study were explained to the patient  and informed  consent  was  ob- tained.  Demographic  data,  past  medical history  and  ocular  history  were  obtained from all patients.

The  worse  eye  was  chosen  based  on the  humphrey  visual  field.  Eyes  with counting  finger  or  worse  vision  were  ex- cluded.  Visual  acuity  using  the  logMar chart  was  recorded  and  autorefraction was done to look at the level of refractive error. Anterior segment examination was performed  and  intraocular  pressure  was measured  before  gonioscopy  was  done. All patients were dilated and examination of  the  retina  was  performed  to  look  for the  presence  of  other  ocular  pathology which  may  affect  the  RNFL  measure- ment.  Optical  coherence  tomography was  done  for  all  patients  with  a  dilated pupil  and  only  those  with  a  good  signal strength were included in the study. Data was  then  collected  and  those  with  in- complete  documentation  were  excluded. Data of the patients were then analyzed.  

RESULTS 

A  total  of  190  eyes  from  190  patients were  included  in  the  study.  There  were 60  patients  in  the  normal  tension  glau- coma  (NTG)  group,  61  patients  for  pri- mary  open  angle  glaucoma  (POAG) group  and  69  patients  for  the  normal (control)  group.  Age  for  all  the  groups were normally distributed (Shapiro-Wilk p >  0.05).  Parametric  statistical  tests  were used  for  the  data  analyses.  All  p-values were 2-sided and were considered statis- tically  significant  when  the  values  were less than 0.05.

Mean  age  for  the POAG  group  was 64.5±10.2  years,  NTG  group  65.0±7.6 years  and  the normal  group  61.6±8.2 years  old  (One  way  ANOVA,  p=0.195). Intraocular  pressure  was  significantly different  between  the  three  groups (p<0.0001)  (Table  1).  Central  corneal thickness of the NTG group was found to be thinnest among the three groups. Sig- nificant  differences  were  noted  between the  CCT  of  the NTG  patients  compared to the control group (p < 0.05) (Table 1).

RNFL  thickness  was  thinner  in  both glaucoma  groups  compared  to  the  con- trol  group  of  patients  (p<0.05)  (Table  2). Both  glaucoma  groups  had  similar  level of severity of disease based on RNFL (p >0.05).

A  significant  Pearson  correlation  coeffi- cient  was  found  in  the  POAG  group  be- tween the CCT and RNFL in the superior quadrant and average thickness (r=0.265, p<0.05 and r=0.417, p<0.05 respectively). No  significant  correlation  was  found between the CCT and RNFL for the NTG group (Table 3, Figures 1, 2). 

DISCUSSION 

Normotensive  glaucoma  was  long  consi- dered  as  a  subset  of  POAG,  sharing many  similar  characteristics  except  for the  elevated  IOP  (Sowka  2005).  In  this study, mean CCT for NTG patients were found to be the thinnest compared to the POAG and the control groups. However, based  on  published  calculations, (Doughty & Zaman 2000) the differences   were too small to influence the IOP mea- surement  in  a  clinically  meaningful  way. Therefore  the  measured  IOP  achieved using  the  'gold  standard'  technique  of Goldmann  tonometry  may  not  differ significantly  from  the  actual  IOP  in  NTG patients.  However,  POAG  patients  were found  to  have  relatively  comparable corneal  thickness  with  the  normal population  supporting  the  fact  that  the disease  is  truly  an  effect  of  an  ageing trabecular meshwork. The value for CCT  in this study however was slightly lower compared  to other  studies  which  mainly used  ultrasound  pachymetry  to  measure the CCT (Bechmann et al. 2000).      

Significant correlation between the CCT and severity of RNFL was only found for the  POAG  group  but  not  for  the NTG group  of  patients.  The  results  obtained were  also  comparable  with  a cross  sec- tional  study  done  using  AGIS  scoring which did not find any significant correla- tion  between  the  corneal  thickness  and severity  of  the disease  in  NTG  patients (Bechmann  et al. 2000).  The  different findings  between  these  two  groups  of patients  may  implicate  the  possibility  of different pathophysiology of the disease.

 Cioffi  and  Liebman  (2002)  proposed possible  relationship  between  a  thin  cor- nea  and  intracellular  matrix  changes  at the  level  of  lamina  cribosa  may  predis- pose  to  glaucomatous  changes.  This idea would be an additional risk factor for this group of patients. Despite controlling the  IOP,  progression  may  seem  to  be inevitable  in  the  presence  of  a  thin  cor- nea. However, this possibility may not be applicable to the NTG patients as similar correlation between the CCT and severity of the disease was not found. Therefore, consideration  of  NTG  as  a  subset  of POAG  has to  be  evaluated  with caution. Although  clinically  the  appearance  on gonioscopy  may  be  similar,  further  eval- uation  at  the  cellular  and  genetic  level may  help  to  answer  these  questions. Other  influential  factors  that  may  contri- bute to the severity of the disease for the NTG  group  were  proposed  by  Doyle (Doyle et al. 2005). Higher association of systemic  disease  in the NTG  group may affect  the ocular  perfusion  indirectly leading  to optic  nerve  damage  in this group of patients (Doyle et al. 2005).

To date, there has been no study done to  look  at  the  association  of  CCT  and severity  of  glaucoma  in  the  NTG  group using the OCT. However, this study  was a  cross  sectional  study  and  the  stability of  the  disease  was  not  analyzed.  As such, the possibility of further RNFL pro- gression  was  not  taken  into  considera- tion.  A  prospective  long  term  study  may give a better understanding of the corre- lation between the corneal thickness and RNFL. 

CONCLUSION 

Significant  correlation  between  the  cor- neal  thickness  and  severity  of  glaucoma is only present in POAG patients but not for  NTG  cases.  Our  results  suggest  that CCT  is  related  to  the  severity  of  POAG- related  visual  loss  thus  measuring  the corneal thickness in this group of patients may  help  to  determine  which  patient would  benefit  from  close  monitoring. Furthermore,  the  old  consideration  of NTG  as  a  subset  of  POAG  may  not  be applicable  with  many  studies  showing dissimilar  detail  characteristics  between those two groups. 

ACKNOWLEDGEMENTS

 The  authors  would  like  to  express  their greatest  gratitude  to  the  Department  of Ophtalmology  for  the  continous  support throughout  the  study,  Prof  Dr  Ropilah Abdul  Rahman,  Dr  Fam  Han  Borr  from Singapore  National  Eye  Centre  and  Dr Nor Fariza Ngah from Hospital Selayang for their critical comments on this study.