diff --git a/pandas/io/parsers/arrow_parser_wrapper.py b/pandas/io/parsers/arrow_parser_wrapper.py
index a7f01e6322755..7fe5ecb0e54c2 100644
--- a/pandas/io/parsers/arrow_parser_wrapper.py
+++ b/pandas/io/parsers/arrow_parser_wrapper.py
@@ -165,7 +165,8 @@ def _finalize_pandas_output(self, frame: DataFrame) -> DataFrame:
                 # The only way self.names is not the same length as number of cols is
                 # if we have int index_col. We should just pad the names(they will get
                 # removed anyways) to expected length then.
-                self.names = list(range(num_cols - len(self.names))) + self.names
+                columns_prefix = [str(x) for x in range(num_cols - len(self.names))]
+                self.names = columns_prefix + self.names
                 multi_index_named = False
             frame.columns = self.names
         # we only need the frame not the names
diff --git a/pandas/tests/io/parser/common/test_index.py b/pandas/tests/io/parser/common/test_index.py
index cdd65223a9c9f..aaa14216bd6d6 100644
--- a/pandas/tests/io/parser/common/test_index.py
+++ b/pandas/tests/io/parser/common/test_index.py
@@ -89,9 +89,6 @@ def test_pass_names_with_index(all_parsers, data, kwargs, expected):
 def test_multi_index_no_level_names(
     request, all_parsers, index_col, using_infer_string
 ):
-    if using_infer_string and all_parsers.engine == "pyarrow":
-        # result should have string columns instead of object dtype
-        request.applymarker(pytest.mark.xfail(reason="TODO(infer_string)"))
     data = """index1,index2,A,B,C,D
 foo,one,2,3,4,5
 foo,two,7,8,9,10