diff --git a/src/calibre/ebooks/oeb/polish/container.py b/src/calibre/ebooks/oeb/polish/container.py
index 2321a56463..f8580235c2 100644
--- a/src/calibre/ebooks/oeb/polish/container.py
+++ b/src/calibre/ebooks/oeb/polish/container.py
@@ -345,12 +345,23 @@ class Container(object):  # {{{
                 self.remove_from_xml(elem)
                 self.dirty(self.opf_name)
         if removed:
+            for spine in self.opf_xpath('//opf:spine'):
+                tocref = spine.attrib.get('toc', None)
+                if tocref and tocref in removed:
+                    spine.attrib.pop('toc', None)
+                    self.dirty(self.opf_name)
+
             for item in self.opf_xpath('//opf:spine/opf:itemref[@idref]'):
                 idref = item.get('idref')
                 if idref in removed:
                     self.remove_from_xml(item)
                     self.dirty(self.opf_name)
 
+            for meta in self.opf_xpath('//opf:meta[@name="cover" and @content]'):
+                if meta.get('content') in removed:
+                    self.remove_from_xml(meta)
+                    self.dirty(self.opf_name)
+
         for item in self.opf_xpath('//opf:guide/opf:reference[@href]'):
             if self.href_to_name(item.get('href'), self.opf_name) == name:
                 self.remove_from_xml(item)
diff --git a/src/calibre/ebooks/oeb/polish/tests/container.py b/src/calibre/ebooks/oeb/polish/tests/container.py
index e65f25fe6f..609766b6fd 100644
--- a/src/calibre/ebooks/oeb/polish/tests/container.py
+++ b/src/calibre/ebooks/oeb/polish/tests/container.py
@@ -57,3 +57,19 @@ class ContainerTests(BaseTest):
             for c in (c1, c2):
                 c.commit(outpath=x)
 
+    def test_file_removal(self):
+        ' Test removal of files from the container '
+        book = get_simple_book()
+        c = get_container(book, tdir=self.tdir)
+        files = ('toc.ncx', 'cover.png', 'titlepage.xhtml')
+        self.assertIn('titlepage.xhtml', {x[0] for x in c.spine_names})
+        self.assertTrue(c.opf_xpath('//opf:meta[@name="cover"]'))
+        for x in files:
+            c.remove_item(x)
+        self.assertIn(c.opf_name, c.dirtied)
+        self.assertNotIn('titlepage.xhtml', {x[0] for x in c.spine_names})
+        self.assertFalse(c.opf_xpath('//opf:meta[@name="cover"]'))
+        raw = c.serialize_item(c.opf_name).decode('utf-8')
+        for x in files:
+            self.assertNotIn(x, raw)
+