diff --git a/modules/rtp_rtcp/source/rtp_sender.cc b/modules/rtp_rtcp/source/rtp_sender.cc
index 5908a23e26..121efea046 100644
--- a/modules/rtp_rtcp/source/rtp_sender.cc
+++ b/modules/rtp_rtcp/source/rtp_sender.cc
@@ -417,9 +417,11 @@ bool RTPSender::SendOutgoingData(FrameType frame_type,
   if (audio_configured_) {
     TRACE_EVENT_ASYNC_STEP1("webrtc", "Audio", rtp_timestamp, "Send", "type",
                             FrameTypeToString(frame_type));
-
+    // The only known way to produce of RTPFragmentationHeader for audio is
+    // to use the AudioCodingModule directly.
+    RTC_DCHECK(fragmentation == nullptr);
     result = audio_->SendAudio(frame_type, payload_type, rtp_timestamp,
-                               payload_data, payload_size, fragmentation);
+                               payload_data, payload_size);
   } else {
     TRACE_EVENT_ASYNC_STEP1("webrtc", "Video", capture_time_ms,
                             "Send", "type", FrameTypeToString(frame_type));
diff --git a/modules/rtp_rtcp/source/rtp_sender_audio.cc b/modules/rtp_rtcp/source/rtp_sender_audio.cc
index cc45a284ef..26108ac062 100644
--- a/modules/rtp_rtcp/source/rtp_sender_audio.cc
+++ b/modules/rtp_rtcp/source/rtp_sender_audio.cc
@@ -119,8 +119,7 @@ bool RTPSenderAudio::SendAudio(FrameType frame_type,
                                int8_t payload_type,
                                uint32_t rtp_timestamp,
                                const uint8_t* payload_data,
-                               size_t payload_size,
-                               const RTPFragmentationHeader* fragmentation) {
+                               size_t payload_size) {
   // From RFC 4733:
   // A source has wide latitude as to how often it sends event updates. A
   // natural interval is the spacing between non-event audio packets. [...]
@@ -223,21 +222,10 @@ bool RTPSenderAudio::SendAudio(FrameType frame_type,
   packet->SetExtension<AudioLevel>(frame_type == kAudioFrameSpeech,
                                    audio_level_dbov);
 
-  if (fragmentation && fragmentation->fragmentationVectorSize > 0) {
-    // Use the fragment info if we have one.
-    uint8_t* payload =
-        packet->AllocatePayload(1 + fragmentation->fragmentationLength[0]);
-    if (!payload)  // Too large payload buffer.
-      return false;
-    payload[0] = fragmentation->fragmentationPlType[0];
-    memcpy(payload + 1, payload_data + fragmentation->fragmentationOffset[0],
-           fragmentation->fragmentationLength[0]);
-  } else {
-    uint8_t* payload = packet->AllocatePayload(payload_size);
-    if (!payload)  // Too large payload buffer.
-      return false;
-    memcpy(payload, payload_data, payload_size);
-  }
+  uint8_t* payload = packet->AllocatePayload(payload_size);
+  if (!payload)  // Too large payload buffer.
+    return false;
+  memcpy(payload, payload_data, payload_size);
 
   if (!rtp_sender_->AssignSequenceNumber(packet.get()))
     return false;
diff --git a/modules/rtp_rtcp/source/rtp_sender_audio.h b/modules/rtp_rtcp/source/rtp_sender_audio.h
index 92c9615ead..16648cd136 100644
--- a/modules/rtp_rtcp/source/rtp_sender_audio.h
+++ b/modules/rtp_rtcp/source/rtp_sender_audio.h
@@ -39,8 +39,7 @@ class RTPSenderAudio {
                  int8_t payload_type,
                  uint32_t capture_timestamp,
                  const uint8_t* payload_data,
-                 size_t payload_size,
-                 const RTPFragmentationHeader* fragmentation);
+                 size_t payload_size);
 
   // Store the audio level in dBov for
   // header-extension-for-audio-level-indication.