diff --git a/asv_bench/benchmarks/index_object.py b/asv_bench/benchmarks/index_object.py
index 454d9ccdda102..7697c3b9d3840 100644
--- a/asv_bench/benchmarks/index_object.py
+++ b/asv_bench/benchmarks/index_object.py
@@ -219,3 +219,22 @@ def time_min(self):
 
     def time_min_trivial(self):
         self.idx_inc.min()
+
+
+class IndexOps(object):
+    goal_time = 0.2
+
+    def setup(self):
+        N = 10000
+        self.ridx = [RangeIndex(i * 100, (i + 1) * 100) for i in range(N)]
+        self.iidx = [idx.astype(int) for idx in self.ridx]
+        self.oidx = [idx.astype(str) for idx in self.iidx]
+
+    def time_concat_range(self):
+        self.ridx[0].append(self.ridx[1:])
+
+    def time_concat_int(self):
+        self.iidx[0].append(self.iidx[1:])
+
+    def time_concat_obj(self):
+        self.oidx[0].append(self.oidx[1:])
diff --git a/pandas/core/dtypes/concat.py b/pandas/core/dtypes/concat.py
index 93993fd0a0cab..4e15aa50e4319 100644
--- a/pandas/core/dtypes/concat.py
+++ b/pandas/core/dtypes/concat.py
@@ -467,6 +467,11 @@ def _concat_datetimetz(to_concat, name=None):
     return to_concat[0]._simple_new(new_values, tz=tz, name=name)
 
 
+def _concat_index_same_dtype(indexes, klass=None):
+    klass = klass if klass is not None else indexes[0].__class__
+    return klass(np.concatenate([x._values for x in indexes]))
+
+
 def _concat_index_asobject(to_concat, name=None):
     """
     concat all inputs as object. DatetimeIndex, TimedeltaIndex and
@@ -581,16 +586,15 @@ def _concat_rangeindex_same_dtype(indexes):
         elif step is None:
             # First non-empty index had only one element
             if obj._start == start:
-                return _concat_index_asobject(indexes)
+                from pandas import Int64Index
+                return _concat_index_same_dtype(indexes, klass=Int64Index)
             step = obj._start - start
 
         non_consecutive = ((step != obj._step and len(obj) > 1) or
                            (next is not None and obj._start != next))
         if non_consecutive:
-            # Int64Index._append_same_dtype([ix.astype(int) for ix in indexes])
-            # would be preferred... but it currently resorts to
-            # _concat_index_asobject anyway.
-            return _concat_index_asobject(indexes)
+            from pandas import Int64Index
+            return _concat_index_same_dtype(indexes, klass=Int64Index)
 
         if step is not None:
             next = obj[-1] + step
diff --git a/pandas/core/indexes/numeric.py b/pandas/core/indexes/numeric.py
index 1f007b1961e06..b0703869948c2 100644
--- a/pandas/core/indexes/numeric.py
+++ b/pandas/core/indexes/numeric.py
@@ -17,6 +17,7 @@
 from pandas.core.indexes.base import (
     Index, InvalidIndexError, _index_shared_docs)
 from pandas.util._decorators import Appender, cache_readonly
+import pandas.core.dtypes.concat as _concat
 import pandas.core.indexes.base as ibase
 
 
@@ -96,6 +97,9 @@ def _assert_safe_casting(cls, data, subarr):
         """
         pass
 
+    def _concat_same_dtype(self, indexes, name):
+        return _concat._concat_index_same_dtype(indexes).rename(name)
+
     @property
     def is_all_dates(self):
         """