diff --git a/src/openmc_mcnp_adapter/parse.py b/src/openmc_mcnp_adapter/parse.py
index 05060e0..dcf9a26 100644
--- a/src/openmc_mcnp_adapter/parse.py
+++ b/src/openmc_mcnp_adapter/parse.py
@@ -377,8 +377,9 @@ def sanitize(section: str) -> str:
     # Remove end-of-line comments
     section = re.sub(r'\$.*$', '', section, flags=re.MULTILINE)
 
-    # Remove comment cards
-    section = re.sub('^[ \t]*?[cC].*?$\n?', '', section, flags=re.MULTILINE)
+    # Remove comment cards: 'c' in first 5 columns followed by at least one
+    # blank, or 'c' as the only character on the line
+    section = re.sub(r'^[ \t]{0,4}[cC](?:[ \t]+.*)?$\n?', '', section, flags=re.MULTILINE)
 
     # Turn continuation lines into single line
     section = re.sub('&.*\n', ' ', section)
diff --git a/tests/test_material.py b/tests/test_material.py
index dc3f7da..a360271 100644
--- a/tests/test_material.py
+++ b/tests/test_material.py
@@ -156,3 +156,27 @@ def test_material_without_keywords():
     nd = m.get_nuclide_densities()
     assert 'U235' in nd and nd['U235'].percent == approx(1.0)
     assert 'U238' in nd and nd['U238'].percent == approx(0.5)
+
+def test_material_with_code_keyword():
+    """Test that keywords starting with 'c' are not treated as comment cards"""
+    # This tests the case where continuation line has keyword starting with 'c'
+    # The 'code=0' line should not be treated as a comment card
+    mcnp_model = textwrap.dedent("""
+        title
+        1   1 -1.0    1
+        2   2 -1.0   -1
+
+        1   px 0.0
+
+        m1  1001.70c 1.0 &
+        code=0
+        m2  1001.70c 1.0
+    """)
+    model = mcnp_str_to_model(mcnp_model)
+    # Should have two distinct materials
+    assert len(model.materials) == 2
+    # Both materials should have H1
+    for mat in model.materials:
+        nd = mat.get_nuclide_densities()
+        assert 'H1' in nd
+        assert nd['H1'].percent == approx(1.0)