diff --git a/examples/merge_anarci_csv.py b/examples/merge_anarci_csv.py
index d6b088e..090fb2a 100755
--- a/examples/merge_anarci_csv.py
+++ b/examples/merge_anarci_csv.py
@@ -6,7 +6,10 @@
 import pandas
 from abnumber import Position
 
-def merge_anarci_csv(input_paths, output_path, scheme):
+def merge_anarci_csv(input_paths, output_path, scheme, add_regions=False):
+    assert input_paths, 'No input files specified'
+
+    # Merge position numbers
     metadata_columns = list()
     position_numbers = set()
     is_heavy_chain = None
@@ -22,27 +25,42 @@ def merge_anarci_csv(input_paths, output_path, scheme):
         for c in df.columns:
             if c not in position_numbers and c not in metadata_columns:
                 metadata_columns.append(c)
-    positions = [Position.from_string(p, 'H' if is_heavy_chain else 'L', scheme) for p in position_numbers]
-    sorted_position_numbers = [p.format(chain_type=False) for p in sorted(positions)] # this uses the correct sorting of Position objects
+    # Use the correct sorting by using Position objects
+    sorted_positions = sorted([Position.from_string(p, 'H' if is_heavy_chain else 'L', scheme) for p in position_numbers])
+    sorted_position_numbers = [p.format(chain_type=False) for p in sorted_positions]
+
+    # Verify that all positions are present
     missing_positions = set(position_numbers) - set(sorted_position_numbers)
     if missing_positions:
         raise ValueError(f'Unexpected error merging positions, missing: {", ".join(missing_positions)}')
+
+    # Merge files
     num = 0
     with open(output_path, 'w') as f:
+        # Write header
+        if add_regions:
+            f.write(','.join(metadata_columns + [p.get_region()+'_'+n for p, n in zip(sorted_positions, sorted_position_numbers)]) + os.linesep)
+        else:
+            f.write(','.join(metadata_columns + sorted_position_numbers) + os.linesep)
+
+        # Write files one by one
         for i, path in enumerate(input_paths):
             df = pandas.read_csv(path)
             df_positions = df.loc[:, ('1' if '1' in df.columns else '2'):].columns.tolist()
             assert df_positions == [c for c in sorted_position_numbers if c in df_positions], \
                 f'Unexpected reordering of positions in CSV file: {sorted_position_numbers}'
+
             df = df.reindex(metadata_columns + sorted_position_numbers, axis=1)
             df.loc[:, sorted_position_numbers].fillna('-', inplace=True)
+            df.to_csv(f, index=False, header=False)
             num += len(df)
-            df.to_csv(f, index=False, header=i == 0)
+
     return num
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser(description='Merge ANARCI CSV files')
     parser.add_argument('--scheme', help='Numbering scheme')
+    parser.add_argument('--add-regions', help='Add region prefixes to header', action='store_true')
     parser.add_argument('input_paths', nargs='+', help='Paths to input CSV files')
     parser.add_argument('output_path', help='Path to output CSV file')
     args = parser.parse_args()
@@ -51,8 +69,12 @@ def merge_anarci_csv(input_paths, output_path, scheme):
         raise ValueError(f'Output file {args.output_path} already exists')
 
     print(f'Merging {len(args.input_paths):,} files...', file=sys.stderr)
-
-    num = merge_anarci_csv(args.input_paths, args.output_path, args.scheme)
-
-    print(f'Merged {num:,} sequences from {len(args.input_paths):,} files', file=sys.stderr)
-    print(f'Saved output to: {args.output_path}', file=sys.stderr)
+    try:
+        num = merge_anarci_csv(args.input_paths, args.output_path, args.scheme, add_regions=args.add_regions)
+        print(f'Merged {num:,} sequences from {len(args.input_paths):,} files', file=sys.stderr)
+        print(f'Saved output to: {args.output_path}', file=sys.stderr)
+    except Exception as e:
+        # Remove output file if an error occurred
+        if os.path.exists(args.output_path):
+            os.remove(args.output_path)
+        raise